diff --git "a/ndarray-cache.json" "b/ndarray-cache.json" new file mode 100644--- /dev/null +++ "b/ndarray-cache.json" @@ -0,0 +1,9965 @@ +{ + "metadata": { + "ParamSize": 869, + "ParamBytes": 2334114368.0, + "BitsPerParam": 4.612736555784706 + }, + "records": [ + { + "dataPath": "params_shard_0.bin", + "format": "raw-shard", + "nbytes": 49250304, + "records": [ + { + "name": "lm_head.q_weight", + "shape": [ + 32064, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 49250304, + "byteOffset": 0 + } + ], + "md5sum": "03f42a0e4b7b45fb1af495013d415f70" + }, + { + "dataPath": "params_shard_1.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.18.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "c26c0bfa7471129355a632bdaceee085" + }, + { + "dataPath": "params_shard_2.bin", + "format": "raw-shard", + "nbytes": 23470080, + "records": [ + { + "name": "lm_head.q_scale", + "shape": [ + 32064, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6156288, + "byteOffset": 0 + }, + { + "name": "model.h.18.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 6156288 + }, + { + "name": "model.h.18.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 6162432 + }, + { + "name": "model.h.18.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 18745344 + }, + { + "name": "model.h.18.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 20318208 + }, + { + "name": "model.h.18.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 23463936 + } + ], + "md5sum": "c76c21adf2577acc8d5081f81635683e" + }, + { + "dataPath": "params_shard_3.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.19.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "41e36be4010532f225910ab2da957df4" + }, + { + "dataPath": "params_shard_4.bin", + "format": "raw-shard", + "nbytes": 33239040, + "records": [ + { + "name": "model.h.18.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 0 + }, + { + "name": "model.h.18.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 14155776 + }, + { + "name": "model.h.19.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 15925248 + }, + { + "name": "model.h.19.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 15931392 + }, + { + "name": "model.h.19.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 28514304 + }, + { + "name": "model.h.19.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 30087168 + }, + { + "name": "model.h.19.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 33232896 + } + ], + "md5sum": "11fe3e24529e5d875b431ad76adbe6c6" + }, + { + "dataPath": "params_shard_5.bin", + "format": "raw-shard", + "nbytes": 21239808, + "records": [ + { + "name": "model.h.19.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 0 + }, + { + "name": "model.h.19.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 4718592 + }, + { + "name": "model.h.19.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 5308416 + }, + { + "name": "model.h.19.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 19464192 + }, + { + "name": "model.h.20.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 21233664 + } + ], + "md5sum": "db5de768c5c55cec34ee218899f883da" + }, + { + "dataPath": "params_shard_6.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.20.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "a7cf0e4fcf6869e0b96ea26b6952735a" + }, + { + "dataPath": "params_shard_7.bin", + "format": "raw-shard", + "nbytes": 22616064, + "records": [ + { + "name": "model.h.20.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 0 + }, + { + "name": "model.h.20.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 12582912 + }, + { + "name": "model.h.20.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 14155776 + }, + { + "name": "model.h.20.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 17301504 + }, + { + "name": "model.h.20.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 17307648 + }, + { + "name": "model.h.20.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 22026240 + } + ], + "md5sum": "f50a0ec8d5b4bc9f25ff5debd51c1210" + }, + { + "dataPath": "params_shard_8.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.21.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "b3731b47d71ec2a2e1a621ce45bce622" + }, + { + "dataPath": "params_shard_9.bin", + "format": "raw-shard", + "nbytes": 33239040, + "records": [ + { + "name": "model.h.20.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 0 + }, + { + "name": "model.h.20.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 14155776 + }, + { + "name": "model.h.21.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 15925248 + }, + { + "name": "model.h.21.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 15931392 + }, + { + "name": "model.h.21.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 28514304 + }, + { + "name": "model.h.21.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 30087168 + }, + { + "name": "model.h.21.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 33232896 + } + ], + "md5sum": "f4b6989d71a06160012b50345da2ab97" + }, + { + "dataPath": "params_shard_10.bin", + "format": "raw-shard", + "nbytes": 21239808, + "records": [ + { + "name": "model.h.21.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 0 + }, + { + "name": "model.h.21.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 4718592 + }, + { + "name": "model.h.21.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 5308416 + }, + { + "name": "model.h.21.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 19464192 + }, + { + "name": "model.h.22.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 21233664 + } + ], + "md5sum": "c7b2fb1b7113e0d9e7ad802025d5a6fc" + }, + { + "dataPath": "params_shard_11.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.22.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "dcc01e341ca10fefd9aac289c3960b97" + }, + { + "dataPath": "params_shard_12.bin", + "format": "raw-shard", + "nbytes": 22616064, + "records": [ + { + "name": "model.h.22.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 0 + }, + { + "name": "model.h.22.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 12582912 + }, + { + "name": "model.h.22.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 14155776 + }, + { + "name": "model.h.22.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 17301504 + }, + { + "name": "model.h.22.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 17307648 + }, + { + "name": "model.h.22.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 22026240 + } + ], + "md5sum": "8cb7f4afa1a3f347d0c443319a527b6a" + }, + { + "dataPath": "params_shard_13.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.23.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "d4fe2733606e5a5c79372261fc28000f" + }, + { + "dataPath": "params_shard_14.bin", + "format": "raw-shard", + "nbytes": 33239040, + "records": [ + { + "name": "model.h.22.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 0 + }, + { + "name": "model.h.22.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 14155776 + }, + { + "name": "model.h.23.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 15925248 + }, + { + "name": "model.h.23.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 15931392 + }, + { + "name": "model.h.23.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 28514304 + }, + { + "name": "model.h.23.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 30087168 + }, + { + "name": "model.h.23.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 33232896 + } + ], + "md5sum": "5d76b6998116de6c8c8d5425a914df0f" + }, + { + "dataPath": "params_shard_15.bin", + "format": "raw-shard", + "nbytes": 21239808, + "records": [ + { + "name": "model.h.23.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 0 + }, + { + "name": "model.h.23.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 4718592 + }, + { + "name": "model.h.23.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 5308416 + }, + { + "name": "model.h.23.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 19464192 + }, + { + "name": "model.h.24.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 21233664 + } + ], + "md5sum": "ef8fe2ffb42f9b91176fa4a9f11bb9de" + }, + { + "dataPath": "params_shard_16.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.24.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "9557dd078ae2cb230face8006995fcf7" + }, + { + "dataPath": "params_shard_17.bin", + "format": "raw-shard", + "nbytes": 22616064, + "records": [ + { + "name": "model.h.24.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 0 + }, + { + "name": "model.h.24.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 12582912 + }, + { + "name": "model.h.24.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 14155776 + }, + { + "name": "model.h.24.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 17301504 + }, + { + "name": "model.h.24.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 17307648 + }, + { + "name": "model.h.24.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 22026240 + } + ], + "md5sum": "3baec0d24d4200875fc06af28c3fb0b8" + }, + { + "dataPath": "params_shard_18.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.25.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "65fbda0ea3fad96582fbcac92f4400fc" + }, + { + "dataPath": "params_shard_19.bin", + "format": "raw-shard", + "nbytes": 33239040, + "records": [ + { + "name": "model.h.24.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 0 + }, + { + "name": "model.h.24.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 14155776 + }, + { + "name": "model.h.25.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 15925248 + }, + { + "name": "model.h.25.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 15931392 + }, + { + "name": "model.h.25.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 28514304 + }, + { + "name": "model.h.25.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 30087168 + }, + { + "name": "model.h.25.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 33232896 + } + ], + "md5sum": "75a0298eea85723412219cc9cb8016a4" + }, + { + "dataPath": "params_shard_20.bin", + "format": "raw-shard", + "nbytes": 21239808, + "records": [ + { + "name": "model.h.25.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 0 + }, + { + "name": "model.h.25.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 4718592 + }, + { + "name": "model.h.25.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 5308416 + }, + { + "name": "model.h.25.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 19464192 + }, + { + "name": "model.h.26.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 21233664 + } + ], + "md5sum": "c6b63b21d6d195d1b184b030c167b1c9" + }, + { + "dataPath": "params_shard_21.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.26.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "324d6fab6d373401073546ed03d39211" + }, + { + "dataPath": "params_shard_22.bin", + "format": "raw-shard", + "nbytes": 22616064, + "records": [ + { + "name": "model.h.26.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 0 + }, + { + "name": "model.h.26.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 12582912 + }, + { + "name": "model.h.26.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 14155776 + }, + { + "name": "model.h.26.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 17301504 + }, + { + "name": "model.h.26.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 17307648 + }, + { + "name": "model.h.26.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 22026240 + } + ], + "md5sum": "8e042a2e177639cb2ff674f3a7266210" + }, + { + "dataPath": "params_shard_23.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.27.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "e92a94d89f861c2c5e9d08f206b7bc1c" + }, + { + "dataPath": "params_shard_24.bin", + "format": "raw-shard", + "nbytes": 33239040, + "records": [ + { + "name": "model.h.26.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 0 + }, + { + "name": "model.h.26.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 14155776 + }, + { + "name": "model.h.27.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 15925248 + }, + { + "name": "model.h.27.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 15931392 + }, + { + "name": "model.h.27.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 28514304 + }, + { + "name": "model.h.27.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 30087168 + }, + { + "name": "model.h.27.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 33232896 + } + ], + "md5sum": "af4e357109ce0dd01c3ea75acb8aaa1a" + }, + { + "dataPath": "params_shard_25.bin", + "format": "raw-shard", + "nbytes": 21239808, + "records": [ + { + "name": "model.h.27.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 0 + }, + { + "name": "model.h.27.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 4718592 + }, + { + "name": "model.h.27.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 5308416 + }, + { + "name": "model.h.27.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 19464192 + }, + { + "name": "model.h.28.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 21233664 + } + ], + "md5sum": "303caa4cee9e70c8bd1066464c2ea849" + }, + { + "dataPath": "params_shard_26.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.28.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "70b2cf47e9e55432dd9088afd8e9a83e" + }, + { + "dataPath": "params_shard_27.bin", + "format": "raw-shard", + "nbytes": 22616064, + "records": [ + { + "name": "model.h.28.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 0 + }, + { + "name": "model.h.28.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 12582912 + }, + { + "name": "model.h.28.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 14155776 + }, + { + "name": "model.h.28.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 17301504 + }, + { + "name": "model.h.28.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 17307648 + }, + { + "name": "model.h.28.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 22026240 + } + ], + "md5sum": "c94a2d330f9cb37c8a61cf90deb0b45f" + }, + { + "dataPath": "params_shard_28.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.29.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "b76a9a27c46ac52bddcd744b605564a9" + }, + { + "dataPath": "params_shard_29.bin", + "format": "raw-shard", + "nbytes": 33239040, + "records": [ + { + "name": "model.h.28.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 0 + }, + { + "name": "model.h.28.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 14155776 + }, + { + "name": "model.h.29.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 15925248 + }, + { + "name": "model.h.29.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 15931392 + }, + { + "name": "model.h.29.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 28514304 + }, + { + "name": "model.h.29.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 30087168 + }, + { + "name": "model.h.29.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 33232896 + } + ], + "md5sum": "fc3908ef1ff4fab5a81adf786477d8b4" + }, + { + "dataPath": "params_shard_30.bin", + "format": "raw-shard", + "nbytes": 21239808, + "records": [ + { + "name": "model.h.29.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 0 + }, + { + "name": "model.h.29.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 4718592 + }, + { + "name": "model.h.29.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 5308416 + }, + { + "name": "model.h.29.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 19464192 + }, + { + "name": "model.h.30.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 21233664 + } + ], + "md5sum": "3df7172b8ec78578295ca75338a0c32e" + }, + { + "dataPath": "params_shard_31.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.30.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "bd5f5975f4b7f21f2554644da9da6f5a" + }, + { + "dataPath": "params_shard_32.bin", + "format": "raw-shard", + "nbytes": 22616064, + "records": [ + { + "name": "model.h.30.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 0 + }, + { + "name": "model.h.30.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 12582912 + }, + { + "name": "model.h.30.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 14155776 + }, + { + "name": "model.h.30.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 17301504 + }, + { + "name": "model.h.30.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 17307648 + }, + { + "name": "model.h.30.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 22026240 + } + ], + "md5sum": "fb63daef7fca12dea591efd21b9d9c14" + }, + { + "dataPath": "params_shard_33.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.31.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "6511803dfb93c3e3ce95b0dd94e1270d" + }, + { + "dataPath": "params_shard_34.bin", + "format": "raw-shard", + "nbytes": 33239040, + "records": [ + { + "name": "model.h.30.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 0 + }, + { + "name": "model.h.30.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 14155776 + }, + { + "name": "model.h.31.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 15925248 + }, + { + "name": "model.h.31.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 15931392 + }, + { + "name": "model.h.31.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 28514304 + }, + { + "name": "model.h.31.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 30087168 + }, + { + "name": "model.h.31.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 33232896 + } + ], + "md5sum": "acd04c251a6a0b8a09fa648724078f9f" + }, + { + "dataPath": "params_shard_35.bin", + "format": "raw-shard", + "nbytes": 49250304, + "records": [ + { + "name": "model.embd.q_weight", + "shape": [ + 32064, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 49250304, + "byteOffset": 0 + } + ], + "md5sum": "424943277254877ab0a7690d3839c6af" + }, + { + "dataPath": "params_shard_36.bin", + "format": "raw-shard", + "nbytes": 27402240, + "records": [ + { + "name": "model.h.31.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 0 + }, + { + "name": "model.h.31.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 4718592 + }, + { + "name": "model.h.31.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 5308416 + }, + { + "name": "model.h.31.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 19464192 + }, + { + "name": "model.norm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 21233664 + }, + { + "name": "model.embd.q_scale", + "shape": [ + 32064, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6156288, + "byteOffset": 21239808 + }, + { + "name": "model.h.0.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 27396096 + } + ], + "md5sum": "6e3c5ead604216089bb2bd02a4329a5a" + }, + { + "dataPath": "params_shard_37.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.0.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "044a1b344b2891ed7f019e9757264f23" + }, + { + "dataPath": "params_shard_38.bin", + "format": "raw-shard", + "nbytes": 22616064, + "records": [ + { + "name": "model.h.0.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 0 + }, + { + "name": "model.h.0.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 12582912 + }, + { + "name": "model.h.0.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 14155776 + }, + { + "name": "model.h.0.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 17301504 + }, + { + "name": "model.h.0.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 17307648 + }, + { + "name": "model.h.0.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 22026240 + } + ], + "md5sum": "203e211d7b41621ac79a6cf68e143d10" + }, + { + "dataPath": "params_shard_39.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.1.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "e85ef9ff37755c08ce31c1aa203216a6" + }, + { + "dataPath": "params_shard_40.bin", + "format": "raw-shard", + "nbytes": 33239040, + "records": [ + { + "name": "model.h.0.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 0 + }, + { + "name": "model.h.0.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 14155776 + }, + { + "name": "model.h.1.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 15925248 + }, + { + "name": "model.h.1.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 15931392 + }, + { + "name": "model.h.1.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 28514304 + }, + { + "name": "model.h.1.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 30087168 + }, + { + "name": "model.h.1.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 33232896 + } + ], + "md5sum": "315bdc92c799dd2d5e4f3d5f56606ab0" + }, + { + "dataPath": "params_shard_41.bin", + "format": "raw-shard", + "nbytes": 21239808, + "records": [ + { + "name": "model.h.1.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 0 + }, + { + "name": "model.h.1.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 4718592 + }, + { + "name": "model.h.1.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 5308416 + }, + { + "name": "model.h.1.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 19464192 + }, + { + "name": "model.h.10.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 21233664 + } + ], + "md5sum": "54074992d6adab3a53bb5113104dc2a2" + }, + { + "dataPath": "params_shard_42.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.10.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "583cf91ed8d73e29c53be3ba8e6b0157" + }, + { + "dataPath": "params_shard_43.bin", + "format": "raw-shard", + "nbytes": 22616064, + "records": [ + { + "name": "model.h.10.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 0 + }, + { + "name": "model.h.10.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 12582912 + }, + { + "name": "model.h.10.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 14155776 + }, + { + "name": "model.h.10.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 17301504 + }, + { + "name": "model.h.10.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 17307648 + }, + { + "name": "model.h.10.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 22026240 + } + ], + "md5sum": "6ec2356dc75920d46ae39ff75a61e327" + }, + { + "dataPath": "params_shard_44.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.11.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "0c8fac4fdb901171c0687163c39bfb24" + }, + { + "dataPath": "params_shard_45.bin", + "format": "raw-shard", + "nbytes": 33239040, + "records": [ + { + "name": "model.h.10.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 0 + }, + { + "name": "model.h.10.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 14155776 + }, + { + "name": "model.h.11.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 15925248 + }, + { + "name": "model.h.11.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 15931392 + }, + { + "name": "model.h.11.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 28514304 + }, + { + "name": "model.h.11.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 30087168 + }, + { + "name": "model.h.11.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 33232896 + } + ], + "md5sum": "f04da33a046700d1fd0c52d1a748b20f" + }, + { + "dataPath": "params_shard_46.bin", + "format": "raw-shard", + "nbytes": 21239808, + "records": [ + { + "name": "model.h.11.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 0 + }, + { + "name": "model.h.11.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 4718592 + }, + { + "name": "model.h.11.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 5308416 + }, + { + "name": "model.h.11.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 19464192 + }, + { + "name": "model.h.12.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 21233664 + } + ], + "md5sum": "94bce31b25ff3751bd949ee2651ff3b0" + }, + { + "dataPath": "params_shard_47.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.12.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "deb5606d247d964e84946eb2acf96f17" + }, + { + "dataPath": "params_shard_48.bin", + "format": "raw-shard", + "nbytes": 22616064, + "records": [ + { + "name": "model.h.12.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 0 + }, + { + "name": "model.h.12.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 12582912 + }, + { + "name": "model.h.12.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 14155776 + }, + { + "name": "model.h.12.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 17301504 + }, + { + "name": "model.h.12.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 17307648 + }, + { + "name": "model.h.12.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 22026240 + } + ], + "md5sum": "147b605c605618800546ef40d9f7dfd9" + }, + { + "dataPath": "params_shard_49.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.13.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "c9e508bb9637f442a61701f39953c947" + }, + { + "dataPath": "params_shard_50.bin", + "format": "raw-shard", + "nbytes": 33239040, + "records": [ + { + "name": "model.h.12.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 0 + }, + { + "name": "model.h.12.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 14155776 + }, + { + "name": "model.h.13.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 15925248 + }, + { + "name": "model.h.13.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 15931392 + }, + { + "name": "model.h.13.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 28514304 + }, + { + "name": "model.h.13.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 30087168 + }, + { + "name": "model.h.13.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 33232896 + } + ], + "md5sum": "54e8325797a3869da3e17f7137347f37" + }, + { + "dataPath": "params_shard_51.bin", + "format": "raw-shard", + "nbytes": 21239808, + "records": [ + { + "name": "model.h.13.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 0 + }, + { + "name": "model.h.13.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 4718592 + }, + { + "name": "model.h.13.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 5308416 + }, + { + "name": "model.h.13.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 19464192 + }, + { + "name": "model.h.14.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 21233664 + } + ], + "md5sum": "d2df7b8d4d0a89dd25e2c890d09ba7ac" + }, + { + "dataPath": "params_shard_52.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.14.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "8f64903fe513470efaea44cb44882554" + }, + { + "dataPath": "params_shard_53.bin", + "format": "raw-shard", + "nbytes": 22616064, + "records": [ + { + "name": "model.h.14.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 0 + }, + { + "name": "model.h.14.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 12582912 + }, + { + "name": "model.h.14.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 14155776 + }, + { + "name": "model.h.14.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 17301504 + }, + { + "name": "model.h.14.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 17307648 + }, + { + "name": "model.h.14.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 22026240 + } + ], + "md5sum": "3aa4555b974bea81381743734646ac6f" + }, + { + "dataPath": "params_shard_54.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.15.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "9bdaa9ffa45dbeba3f15230433248a45" + }, + { + "dataPath": "params_shard_55.bin", + "format": "raw-shard", + "nbytes": 33239040, + "records": [ + { + "name": "model.h.14.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 0 + }, + { + "name": "model.h.14.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 14155776 + }, + { + "name": "model.h.15.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 15925248 + }, + { + "name": "model.h.15.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 15931392 + }, + { + "name": "model.h.15.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 28514304 + }, + { + "name": "model.h.15.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 30087168 + }, + { + "name": "model.h.15.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 33232896 + } + ], + "md5sum": "a4838335ebbe75771cf31aee80e9386b" + }, + { + "dataPath": "params_shard_56.bin", + "format": "raw-shard", + "nbytes": 21239808, + "records": [ + { + "name": "model.h.15.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 0 + }, + { + "name": "model.h.15.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 4718592 + }, + { + "name": "model.h.15.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 5308416 + }, + { + "name": "model.h.15.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 19464192 + }, + { + "name": "model.h.16.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 21233664 + } + ], + "md5sum": "54597a723cf21d0eef770522979d21ba" + }, + { + "dataPath": "params_shard_57.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.16.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "55c07771f9d95822295047a776e4e716" + }, + { + "dataPath": "params_shard_58.bin", + "format": "raw-shard", + "nbytes": 22616064, + "records": [ + { + "name": "model.h.16.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 0 + }, + { + "name": "model.h.16.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 12582912 + }, + { + "name": "model.h.16.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 14155776 + }, + { + "name": "model.h.16.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 17301504 + }, + { + "name": "model.h.16.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 17307648 + }, + { + "name": "model.h.16.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 22026240 + } + ], + "md5sum": "e292be869f50f3f22711ede0efca8e99" + }, + { + "dataPath": "params_shard_59.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.17.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "fc5c9bf4993163454cac074473ec84ff" + }, + { + "dataPath": "params_shard_60.bin", + "format": "raw-shard", + "nbytes": 33239040, + "records": [ + { + "name": "model.h.16.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 0 + }, + { + "name": "model.h.16.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 14155776 + }, + { + "name": "model.h.17.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 15925248 + }, + { + "name": "model.h.17.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 15931392 + }, + { + "name": "model.h.17.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 28514304 + }, + { + "name": "model.h.17.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 30087168 + }, + { + "name": "model.h.17.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 33232896 + } + ], + "md5sum": "35b09aad431916ab9986914db9feaa15" + }, + { + "dataPath": "params_shard_61.bin", + "format": "raw-shard", + "nbytes": 26548224, + "records": [ + { + "name": "model.h.17.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 0 + }, + { + "name": "model.h.17.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 4718592 + }, + { + "name": "model.h.17.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 5308416 + }, + { + "name": "model.h.17.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 19464192 + }, + { + "name": "model.h.18.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 21233664 + }, + { + "name": "model.h.18.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 25952256 + }, + { + "name": "model.h.2.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 26542080 + } + ], + "md5sum": "d612bf0a65838fe9cd7c5338b0ab2d1a" + }, + { + "dataPath": "params_shard_62.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.2.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "a334b0a59263051527b3959507c2059f" + }, + { + "dataPath": "params_shard_63.bin", + "format": "raw-shard", + "nbytes": 22616064, + "records": [ + { + "name": "model.h.2.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 0 + }, + { + "name": "model.h.2.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 12582912 + }, + { + "name": "model.h.2.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 14155776 + }, + { + "name": "model.h.2.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 17301504 + }, + { + "name": "model.h.2.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 17307648 + }, + { + "name": "model.h.2.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 22026240 + } + ], + "md5sum": "ff92efa38feb0013fbeeeb24a6beb6a2" + }, + { + "dataPath": "params_shard_64.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.3.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "1030ee4948f7cc79712705800b13e010" + }, + { + "dataPath": "params_shard_65.bin", + "format": "raw-shard", + "nbytes": 33239040, + "records": [ + { + "name": "model.h.2.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 0 + }, + { + "name": "model.h.2.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 14155776 + }, + { + "name": "model.h.3.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 15925248 + }, + { + "name": "model.h.3.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 15931392 + }, + { + "name": "model.h.3.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 28514304 + }, + { + "name": "model.h.3.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 30087168 + }, + { + "name": "model.h.3.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 33232896 + } + ], + "md5sum": "1d2e26f38ce827dce9e614d53fbae96b" + }, + { + "dataPath": "params_shard_66.bin", + "format": "raw-shard", + "nbytes": 21239808, + "records": [ + { + "name": "model.h.3.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 0 + }, + { + "name": "model.h.3.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 4718592 + }, + { + "name": "model.h.3.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 5308416 + }, + { + "name": "model.h.3.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 19464192 + }, + { + "name": "model.h.4.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 21233664 + } + ], + "md5sum": "5d4f870aa25b56a966e21e557b6a5689" + }, + { + "dataPath": "params_shard_67.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.4.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "b8512fae8f93fbd9f4f7f73337311990" + }, + { + "dataPath": "params_shard_68.bin", + "format": "raw-shard", + "nbytes": 22616064, + "records": [ + { + "name": "model.h.4.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 0 + }, + { + "name": "model.h.4.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 12582912 + }, + { + "name": "model.h.4.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 14155776 + }, + { + "name": "model.h.4.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 17301504 + }, + { + "name": "model.h.4.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 17307648 + }, + { + "name": "model.h.4.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 22026240 + } + ], + "md5sum": "dca262e67f5875824b6cb19b77b87f2b" + }, + { + "dataPath": "params_shard_69.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.5.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "95e02febe339e87fd2dc66e8720d5724" + }, + { + "dataPath": "params_shard_70.bin", + "format": "raw-shard", + "nbytes": 33239040, + "records": [ + { + "name": "model.h.4.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 0 + }, + { + "name": "model.h.4.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 14155776 + }, + { + "name": "model.h.5.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 15925248 + }, + { + "name": "model.h.5.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 15931392 + }, + { + "name": "model.h.5.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 28514304 + }, + { + "name": "model.h.5.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 30087168 + }, + { + "name": "model.h.5.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 33232896 + } + ], + "md5sum": "9ffc11f2725ba1a6a50397edb5970b7b" + }, + { + "dataPath": "params_shard_71.bin", + "format": "raw-shard", + "nbytes": 21239808, + "records": [ + { + "name": "model.h.5.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 0 + }, + { + "name": "model.h.5.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 4718592 + }, + { + "name": "model.h.5.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 5308416 + }, + { + "name": "model.h.5.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 19464192 + }, + { + "name": "model.h.6.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 21233664 + } + ], + "md5sum": "b82024bf6121748baaa0b2e15f45292a" + }, + { + "dataPath": "params_shard_72.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.6.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "74351377410f94f98f638d90086a3ba3" + }, + { + "dataPath": "params_shard_73.bin", + "format": "raw-shard", + "nbytes": 22616064, + "records": [ + { + "name": "model.h.6.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 0 + }, + { + "name": "model.h.6.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 12582912 + }, + { + "name": "model.h.6.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 14155776 + }, + { + "name": "model.h.6.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 17301504 + }, + { + "name": "model.h.6.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 17307648 + }, + { + "name": "model.h.6.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 22026240 + } + ], + "md5sum": "8fcac7f6ef447f7857682429ee194f8c" + }, + { + "dataPath": "params_shard_74.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.7.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "92d2cc1b281e6bb28f465b49ff9b208c" + }, + { + "dataPath": "params_shard_75.bin", + "format": "raw-shard", + "nbytes": 33239040, + "records": [ + { + "name": "model.h.6.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 0 + }, + { + "name": "model.h.6.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 14155776 + }, + { + "name": "model.h.7.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 15925248 + }, + { + "name": "model.h.7.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 15931392 + }, + { + "name": "model.h.7.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 28514304 + }, + { + "name": "model.h.7.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 30087168 + }, + { + "name": "model.h.7.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 33232896 + } + ], + "md5sum": "5db69c637c72f53238fa9c6f55f2d4ef" + }, + { + "dataPath": "params_shard_76.bin", + "format": "raw-shard", + "nbytes": 21239808, + "records": [ + { + "name": "model.h.7.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 0 + }, + { + "name": "model.h.7.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 4718592 + }, + { + "name": "model.h.7.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 5308416 + }, + { + "name": "model.h.7.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 19464192 + }, + { + "name": "model.h.8.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 21233664 + } + ], + "md5sum": "18aac3937281d4613c4bebe4fef0e9f5" + }, + { + "dataPath": "params_shard_77.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.8.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "e6cfafe2fc17539ef7a158a9c123d35c" + }, + { + "dataPath": "params_shard_78.bin", + "format": "raw-shard", + "nbytes": 22616064, + "records": [ + { + "name": "model.h.8.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 0 + }, + { + "name": "model.h.8.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 12582912 + }, + { + "name": "model.h.8.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 14155776 + }, + { + "name": "model.h.8.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 17301504 + }, + { + "name": "model.h.8.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 17307648 + }, + { + "name": "model.h.8.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 22026240 + } + ], + "md5sum": "d095dac880afc3e18c3ce261f3fa747e" + }, + { + "dataPath": "params_shard_79.bin", + "format": "raw-shard", + "nbytes": 25165824, + "records": [ + { + "name": "model.h.9.mlp.gate_up_proj.q_weight", + "shape": [ + 16384, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25165824, + "byteOffset": 0 + } + ], + "md5sum": "ced88ec5fc5b74c9a3131e2cbc474bc9" + }, + { + "dataPath": "params_shard_80.bin", + "format": "raw-shard", + "nbytes": 33239040, + "records": [ + { + "name": "model.h.8.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 0 + }, + { + "name": "model.h.8.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 14155776 + }, + { + "name": "model.h.9.ln.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 15925248 + }, + { + "name": "model.h.9.mlp.down_proj.q_weight", + "shape": [ + 3072, + 1024 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 12582912, + "byteOffset": 15931392 + }, + { + "name": "model.h.9.mlp.down_proj.q_scale", + "shape": [ + 3072, + 256 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1572864, + "byteOffset": 28514304 + }, + { + "name": "model.h.9.mlp.gate_up_proj.q_scale", + "shape": [ + 16384, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 3145728, + "byteOffset": 30087168 + }, + { + "name": "model.h.9.post_attention_layernorm.weight", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 33232896 + } + ], + "md5sum": "62a16d6f9c5485e7c09d371f33dd08d3" + }, + { + "dataPath": "params_shard_81.bin", + "format": "raw-shard", + "nbytes": 32262720, + "records": [ + { + "name": "model.h.9.mixer.out_proj.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 0 + }, + { + "name": "model.h.9.mixer.out_proj.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 4718592 + }, + { + "name": "model.h.9.mixer.qkv_proj.q_weight", + "shape": [ + 9216, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14155776, + "byteOffset": 5308416 + }, + { + "name": "model.h.9.mixer.qkv_proj.q_scale", + "shape": [ + 9216, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1769472, + "byteOffset": 19464192 + }, + { + "name": "vision_embed_tokens.glb_GN", + "shape": [ + 1, + 1, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8192, + "byteOffset": 21233664 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.embeddings.class_embedding", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 21241856 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.embeddings.patch_embedding.weight", + "shape": [ + 1024, + 3, + 14, + 14 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1204224, + "byteOffset": 21243904 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.embeddings.position_embedding.q_weight", + "shape": [ + 577, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 295424, + "byteOffset": 22448128 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.embeddings.position_embedding.q_scale", + "shape": [ + 577, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 36928, + "byteOffset": 22743552 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.0.layer_norm1.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 22780480 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.0.layer_norm1.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 22782528 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.0.layer_norm2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 22784576 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.0.layer_norm2.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 22786624 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.0.mlp.fc1.bias", + "shape": [ + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8192, + "byteOffset": 22788672 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.0.mlp.fc1.q_weight", + "shape": [ + 4096, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 22796864 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.0.mlp.fc1.q_scale", + "shape": [ + 4096, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 24894016 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.0.mlp.fc2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 25156160 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.0.mlp.fc2.q_weight", + "shape": [ + 1024, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 25158208 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.0.mlp.fc2.q_scale", + "shape": [ + 1024, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 27255360 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.0.self_attn.k_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 27517504 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.0.self_attn.k_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 27519552 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.0.self_attn.k_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 28043840 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.0.self_attn.out_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 28109376 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.0.self_attn.out_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 28111424 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.0.self_attn.out_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 28635712 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.0.self_attn.q_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 28701248 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.0.self_attn.q_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 28703296 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.0.self_attn.q_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 29227584 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.0.self_attn.v_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 29293120 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.0.self_attn.v_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 29295168 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.0.self_attn.v_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 29819456 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.1.layer_norm1.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 29884992 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.1.layer_norm1.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 29887040 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.1.layer_norm2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 29889088 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.1.layer_norm2.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 29891136 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.1.mlp.fc1.bias", + "shape": [ + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8192, + "byteOffset": 29893184 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.1.mlp.fc1.q_weight", + "shape": [ + 4096, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 29901376 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.1.mlp.fc1.q_scale", + "shape": [ + 4096, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 31998528 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.1.mlp.fc2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 32260672 + } + ], + "md5sum": "f31749651cdc2e1f7cb48008e11e163a" + }, + { + "dataPath": "params_shard_82.bin", + "format": "raw-shard", + "nbytes": 33161216, + "records": [ + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.1.mlp.fc2.q_weight", + "shape": [ + 1024, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 0 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.1.mlp.fc2.q_scale", + "shape": [ + 1024, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 2097152 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.1.self_attn.k_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 2359296 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.1.self_attn.k_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 2361344 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.1.self_attn.k_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 2885632 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.1.self_attn.out_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 2951168 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.1.self_attn.out_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 2953216 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.1.self_attn.out_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 3477504 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.1.self_attn.q_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 3543040 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.1.self_attn.q_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 3545088 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.1.self_attn.q_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 4069376 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.1.self_attn.v_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 4134912 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.1.self_attn.v_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 4136960 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.1.self_attn.v_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 4661248 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.10.layer_norm1.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 4726784 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.10.layer_norm1.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 4728832 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.10.layer_norm2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 4730880 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.10.layer_norm2.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 4732928 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.10.mlp.fc1.bias", + "shape": [ + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8192, + "byteOffset": 4734976 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.10.mlp.fc1.q_weight", + "shape": [ + 4096, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 4743168 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.10.mlp.fc1.q_scale", + "shape": [ + 4096, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 6840320 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.10.mlp.fc2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 7102464 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.10.mlp.fc2.q_weight", + "shape": [ + 1024, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 7104512 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.10.mlp.fc2.q_scale", + "shape": [ + 1024, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 9201664 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.10.self_attn.k_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 9463808 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.10.self_attn.k_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 9465856 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.10.self_attn.k_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 9990144 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.10.self_attn.out_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 10055680 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.10.self_attn.out_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 10057728 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.10.self_attn.out_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 10582016 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.10.self_attn.q_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 10647552 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.10.self_attn.q_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 10649600 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.10.self_attn.q_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 11173888 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.10.self_attn.v_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 11239424 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.10.self_attn.v_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 11241472 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.10.self_attn.v_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 11765760 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.11.layer_norm1.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 11831296 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.11.layer_norm1.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 11833344 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.11.layer_norm2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 11835392 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.11.layer_norm2.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 11837440 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.11.mlp.fc1.bias", + "shape": [ + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8192, + "byteOffset": 11839488 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.11.mlp.fc1.q_weight", + "shape": [ + 4096, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 11847680 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.11.mlp.fc1.q_scale", + "shape": [ + 4096, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 13944832 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.11.mlp.fc2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 14206976 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.11.mlp.fc2.q_weight", + "shape": [ + 1024, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 14209024 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.11.mlp.fc2.q_scale", + "shape": [ + 1024, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 16306176 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.11.self_attn.k_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 16568320 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.11.self_attn.k_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 16570368 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.11.self_attn.k_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 17094656 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.11.self_attn.out_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 17160192 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.11.self_attn.out_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 17162240 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.11.self_attn.out_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 17686528 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.11.self_attn.q_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 17752064 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.11.self_attn.q_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 17754112 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.11.self_attn.q_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 18278400 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.11.self_attn.v_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 18343936 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.11.self_attn.v_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 18345984 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.11.self_attn.v_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 18870272 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.12.layer_norm1.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 18935808 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.12.layer_norm1.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 18937856 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.12.layer_norm2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 18939904 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.12.layer_norm2.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 18941952 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.12.mlp.fc1.bias", + "shape": [ + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8192, + "byteOffset": 18944000 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.12.mlp.fc1.q_weight", + "shape": [ + 4096, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 18952192 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.12.mlp.fc1.q_scale", + "shape": [ + 4096, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 21049344 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.12.mlp.fc2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 21311488 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.12.mlp.fc2.q_weight", + "shape": [ + 1024, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 21313536 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.12.mlp.fc2.q_scale", + "shape": [ + 1024, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 23410688 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.12.self_attn.k_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 23672832 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.12.self_attn.k_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 23674880 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.12.self_attn.k_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 24199168 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.12.self_attn.out_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 24264704 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.12.self_attn.out_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 24266752 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.12.self_attn.out_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 24791040 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.12.self_attn.q_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 24856576 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.12.self_attn.q_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 24858624 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.12.self_attn.q_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 25382912 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.12.self_attn.v_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 25448448 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.12.self_attn.v_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 25450496 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.12.self_attn.v_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 25974784 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.13.layer_norm1.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 26040320 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.13.layer_norm1.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 26042368 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.13.layer_norm2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 26044416 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.13.layer_norm2.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 26046464 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.13.mlp.fc1.bias", + "shape": [ + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8192, + "byteOffset": 26048512 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.13.mlp.fc1.q_weight", + "shape": [ + 4096, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 26056704 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.13.mlp.fc1.q_scale", + "shape": [ + 4096, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 28153856 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.13.mlp.fc2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 28416000 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.13.mlp.fc2.q_weight", + "shape": [ + 1024, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 28418048 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.13.mlp.fc2.q_scale", + "shape": [ + 1024, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 30515200 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.13.self_attn.k_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 30777344 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.13.self_attn.k_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 30779392 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.13.self_attn.k_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 31303680 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.13.self_attn.out_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 31369216 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.13.self_attn.out_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 31371264 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.13.self_attn.out_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 31895552 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.13.self_attn.q_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 31961088 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.13.self_attn.q_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 31963136 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.13.self_attn.q_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 32487424 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.13.self_attn.v_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 32552960 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.13.self_attn.v_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 32555008 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.13.self_attn.v_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 33079296 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.14.layer_norm1.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 33144832 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.14.layer_norm1.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 33146880 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.14.layer_norm2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 33148928 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.14.layer_norm2.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 33150976 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.14.mlp.fc1.bias", + "shape": [ + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8192, + "byteOffset": 33153024 + } + ], + "md5sum": "f08b634e46a63d50279fc8be715cd065" + }, + { + "dataPath": "params_shard_83.bin", + "format": "raw-shard", + "nbytes": 33140736, + "records": [ + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.14.mlp.fc1.q_weight", + "shape": [ + 4096, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 0 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.14.mlp.fc1.q_scale", + "shape": [ + 4096, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 2097152 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.14.mlp.fc2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 2359296 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.14.mlp.fc2.q_weight", + "shape": [ + 1024, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 2361344 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.14.mlp.fc2.q_scale", + "shape": [ + 1024, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 4458496 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.14.self_attn.k_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 4720640 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.14.self_attn.k_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 4722688 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.14.self_attn.k_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 5246976 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.14.self_attn.out_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 5312512 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.14.self_attn.out_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 5314560 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.14.self_attn.out_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 5838848 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.14.self_attn.q_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 5904384 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.14.self_attn.q_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 5906432 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.14.self_attn.q_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 6430720 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.14.self_attn.v_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 6496256 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.14.self_attn.v_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 6498304 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.14.self_attn.v_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 7022592 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.15.layer_norm1.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 7088128 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.15.layer_norm1.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 7090176 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.15.layer_norm2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 7092224 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.15.layer_norm2.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 7094272 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.15.mlp.fc1.bias", + "shape": [ + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8192, + "byteOffset": 7096320 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.15.mlp.fc1.q_weight", + "shape": [ + 4096, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 7104512 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.15.mlp.fc1.q_scale", + "shape": [ + 4096, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 9201664 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.15.mlp.fc2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 9463808 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.15.mlp.fc2.q_weight", + "shape": [ + 1024, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 9465856 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.15.mlp.fc2.q_scale", + "shape": [ + 1024, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 11563008 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.15.self_attn.k_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 11825152 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.15.self_attn.k_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 11827200 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.15.self_attn.k_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 12351488 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.15.self_attn.out_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 12417024 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.15.self_attn.out_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 12419072 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.15.self_attn.out_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 12943360 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.15.self_attn.q_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 13008896 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.15.self_attn.q_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 13010944 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.15.self_attn.q_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 13535232 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.15.self_attn.v_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 13600768 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.15.self_attn.v_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 13602816 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.15.self_attn.v_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 14127104 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.16.layer_norm1.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 14192640 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.16.layer_norm1.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 14194688 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.16.layer_norm2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 14196736 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.16.layer_norm2.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 14198784 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.16.mlp.fc1.bias", + "shape": [ + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8192, + "byteOffset": 14200832 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.16.mlp.fc1.q_weight", + "shape": [ + 4096, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 14209024 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.16.mlp.fc1.q_scale", + "shape": [ + 4096, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 16306176 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.16.mlp.fc2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 16568320 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.16.mlp.fc2.q_weight", + "shape": [ + 1024, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 16570368 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.16.mlp.fc2.q_scale", + "shape": [ + 1024, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 18667520 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.16.self_attn.k_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 18929664 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.16.self_attn.k_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 18931712 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.16.self_attn.k_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 19456000 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.16.self_attn.out_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 19521536 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.16.self_attn.out_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 19523584 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.16.self_attn.out_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 20047872 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.16.self_attn.q_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 20113408 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.16.self_attn.q_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 20115456 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.16.self_attn.q_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 20639744 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.16.self_attn.v_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 20705280 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.16.self_attn.v_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 20707328 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.16.self_attn.v_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 21231616 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.17.layer_norm1.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 21297152 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.17.layer_norm1.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 21299200 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.17.layer_norm2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 21301248 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.17.layer_norm2.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 21303296 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.17.mlp.fc1.bias", + "shape": [ + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8192, + "byteOffset": 21305344 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.17.mlp.fc1.q_weight", + "shape": [ + 4096, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 21313536 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.17.mlp.fc1.q_scale", + "shape": [ + 4096, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 23410688 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.17.mlp.fc2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 23672832 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.17.mlp.fc2.q_weight", + "shape": [ + 1024, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 23674880 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.17.mlp.fc2.q_scale", + "shape": [ + 1024, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 25772032 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.17.self_attn.k_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 26034176 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.17.self_attn.k_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 26036224 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.17.self_attn.k_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 26560512 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.17.self_attn.out_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 26626048 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.17.self_attn.out_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 26628096 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.17.self_attn.out_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 27152384 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.17.self_attn.q_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 27217920 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.17.self_attn.q_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 27219968 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.17.self_attn.q_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 27744256 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.17.self_attn.v_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 27809792 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.17.self_attn.v_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 27811840 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.17.self_attn.v_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 28336128 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.18.layer_norm1.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 28401664 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.18.layer_norm1.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 28403712 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.18.layer_norm2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 28405760 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.18.layer_norm2.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 28407808 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.18.mlp.fc1.bias", + "shape": [ + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8192, + "byteOffset": 28409856 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.18.mlp.fc1.q_weight", + "shape": [ + 4096, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 28418048 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.18.mlp.fc1.q_scale", + "shape": [ + 4096, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 30515200 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.18.mlp.fc2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 30777344 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.18.mlp.fc2.q_weight", + "shape": [ + 1024, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 30779392 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.18.mlp.fc2.q_scale", + "shape": [ + 1024, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 32876544 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.18.self_attn.k_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 33138688 + } + ], + "md5sum": "a0e8309f607c57253603f2ffd504cce4" + }, + { + "dataPath": "params_shard_84.bin", + "format": "raw-shard", + "nbytes": 33161216, + "records": [ + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.18.self_attn.k_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 0 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.18.self_attn.k_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 524288 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.18.self_attn.out_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 589824 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.18.self_attn.out_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 591872 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.18.self_attn.out_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 1116160 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.18.self_attn.q_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 1181696 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.18.self_attn.q_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 1183744 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.18.self_attn.q_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 1708032 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.18.self_attn.v_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 1773568 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.18.self_attn.v_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 1775616 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.18.self_attn.v_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 2299904 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.19.layer_norm1.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 2365440 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.19.layer_norm1.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 2367488 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.19.layer_norm2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 2369536 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.19.layer_norm2.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 2371584 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.19.mlp.fc1.bias", + "shape": [ + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8192, + "byteOffset": 2373632 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.19.mlp.fc1.q_weight", + "shape": [ + 4096, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 2381824 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.19.mlp.fc1.q_scale", + "shape": [ + 4096, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 4478976 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.19.mlp.fc2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 4741120 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.19.mlp.fc2.q_weight", + "shape": [ + 1024, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 4743168 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.19.mlp.fc2.q_scale", + "shape": [ + 1024, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 6840320 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.19.self_attn.k_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 7102464 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.19.self_attn.k_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 7104512 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.19.self_attn.k_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 7628800 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.19.self_attn.out_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 7694336 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.19.self_attn.out_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 7696384 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.19.self_attn.out_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 8220672 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.19.self_attn.q_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 8286208 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.19.self_attn.q_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 8288256 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.19.self_attn.q_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 8812544 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.19.self_attn.v_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 8878080 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.19.self_attn.v_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 8880128 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.19.self_attn.v_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 9404416 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.2.layer_norm1.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 9469952 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.2.layer_norm1.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 9472000 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.2.layer_norm2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 9474048 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.2.layer_norm2.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 9476096 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.2.mlp.fc1.bias", + "shape": [ + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8192, + "byteOffset": 9478144 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.2.mlp.fc1.q_weight", + "shape": [ + 4096, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 9486336 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.2.mlp.fc1.q_scale", + "shape": [ + 4096, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 11583488 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.2.mlp.fc2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 11845632 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.2.mlp.fc2.q_weight", + "shape": [ + 1024, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 11847680 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.2.mlp.fc2.q_scale", + "shape": [ + 1024, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 13944832 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.2.self_attn.k_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 14206976 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.2.self_attn.k_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 14209024 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.2.self_attn.k_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 14733312 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.2.self_attn.out_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 14798848 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.2.self_attn.out_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 14800896 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.2.self_attn.out_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 15325184 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.2.self_attn.q_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 15390720 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.2.self_attn.q_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 15392768 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.2.self_attn.q_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 15917056 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.2.self_attn.v_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 15982592 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.2.self_attn.v_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 15984640 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.2.self_attn.v_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 16508928 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.20.layer_norm1.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 16574464 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.20.layer_norm1.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 16576512 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.20.layer_norm2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 16578560 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.20.layer_norm2.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 16580608 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.20.mlp.fc1.bias", + "shape": [ + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8192, + "byteOffset": 16582656 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.20.mlp.fc1.q_weight", + "shape": [ + 4096, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 16590848 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.20.mlp.fc1.q_scale", + "shape": [ + 4096, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 18688000 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.20.mlp.fc2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 18950144 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.20.mlp.fc2.q_weight", + "shape": [ + 1024, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 18952192 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.20.mlp.fc2.q_scale", + "shape": [ + 1024, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 21049344 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.20.self_attn.k_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 21311488 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.20.self_attn.k_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 21313536 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.20.self_attn.k_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 21837824 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.20.self_attn.out_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 21903360 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.20.self_attn.out_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 21905408 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.20.self_attn.out_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 22429696 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.20.self_attn.q_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 22495232 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.20.self_attn.q_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 22497280 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.20.self_attn.q_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 23021568 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.20.self_attn.v_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 23087104 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.20.self_attn.v_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 23089152 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.20.self_attn.v_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 23613440 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.21.layer_norm1.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 23678976 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.21.layer_norm1.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 23681024 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.21.layer_norm2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 23683072 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.21.layer_norm2.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 23685120 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.21.mlp.fc1.bias", + "shape": [ + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8192, + "byteOffset": 23687168 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.21.mlp.fc1.q_weight", + "shape": [ + 4096, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 23695360 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.21.mlp.fc1.q_scale", + "shape": [ + 4096, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 25792512 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.21.mlp.fc2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 26054656 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.21.mlp.fc2.q_weight", + "shape": [ + 1024, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 26056704 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.21.mlp.fc2.q_scale", + "shape": [ + 1024, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 28153856 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.21.self_attn.k_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 28416000 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.21.self_attn.k_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 28418048 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.21.self_attn.k_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 28942336 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.21.self_attn.out_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 29007872 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.21.self_attn.out_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 29009920 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.21.self_attn.out_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 29534208 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.21.self_attn.q_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 29599744 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.21.self_attn.q_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 29601792 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.21.self_attn.q_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 30126080 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.21.self_attn.v_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 30191616 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.21.self_attn.v_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 30193664 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.21.self_attn.v_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 30717952 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.22.layer_norm1.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 30783488 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.22.layer_norm1.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 30785536 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.22.layer_norm2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 30787584 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.22.layer_norm2.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 30789632 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.22.mlp.fc1.bias", + "shape": [ + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8192, + "byteOffset": 30791680 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.22.mlp.fc1.q_weight", + "shape": [ + 4096, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 30799872 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.22.mlp.fc1.q_scale", + "shape": [ + 4096, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 32897024 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.22.mlp.fc2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 33159168 + } + ], + "md5sum": "10ad505e9ab90b00640b5382d9bb5582" + }, + { + "dataPath": "params_shard_85.bin", + "format": "raw-shard", + "nbytes": 33161216, + "records": [ + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.22.mlp.fc2.q_weight", + "shape": [ + 1024, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 0 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.22.mlp.fc2.q_scale", + "shape": [ + 1024, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 2097152 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.22.self_attn.k_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 2359296 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.22.self_attn.k_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 2361344 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.22.self_attn.k_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 2885632 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.22.self_attn.out_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 2951168 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.22.self_attn.out_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 2953216 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.22.self_attn.out_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 3477504 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.22.self_attn.q_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 3543040 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.22.self_attn.q_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 3545088 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.22.self_attn.q_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 4069376 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.22.self_attn.v_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 4134912 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.22.self_attn.v_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 4136960 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.22.self_attn.v_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 4661248 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.23.layer_norm1.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 4726784 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.23.layer_norm1.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 4728832 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.23.layer_norm2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 4730880 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.23.layer_norm2.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 4732928 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.23.mlp.fc1.bias", + "shape": [ + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8192, + "byteOffset": 4734976 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.23.mlp.fc1.q_weight", + "shape": [ + 4096, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 4743168 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.23.mlp.fc1.q_scale", + "shape": [ + 4096, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 6840320 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.23.mlp.fc2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 7102464 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.23.mlp.fc2.q_weight", + "shape": [ + 1024, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 7104512 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.23.mlp.fc2.q_scale", + "shape": [ + 1024, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 9201664 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.23.self_attn.k_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 9463808 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.23.self_attn.k_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 9465856 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.23.self_attn.k_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 9990144 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.23.self_attn.out_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 10055680 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.23.self_attn.out_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 10057728 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.23.self_attn.out_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 10582016 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.23.self_attn.q_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 10647552 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.23.self_attn.q_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 10649600 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.23.self_attn.q_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 11173888 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.23.self_attn.v_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 11239424 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.23.self_attn.v_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 11241472 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.23.self_attn.v_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 11765760 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.3.layer_norm1.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 11831296 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.3.layer_norm1.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 11833344 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.3.layer_norm2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 11835392 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.3.layer_norm2.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 11837440 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.3.mlp.fc1.bias", + "shape": [ + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8192, + "byteOffset": 11839488 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.3.mlp.fc1.q_weight", + "shape": [ + 4096, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 11847680 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.3.mlp.fc1.q_scale", + "shape": [ + 4096, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 13944832 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.3.mlp.fc2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 14206976 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.3.mlp.fc2.q_weight", + "shape": [ + 1024, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 14209024 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.3.mlp.fc2.q_scale", + "shape": [ + 1024, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 16306176 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.3.self_attn.k_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 16568320 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.3.self_attn.k_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 16570368 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.3.self_attn.k_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 17094656 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.3.self_attn.out_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 17160192 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.3.self_attn.out_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 17162240 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.3.self_attn.out_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 17686528 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.3.self_attn.q_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 17752064 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.3.self_attn.q_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 17754112 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.3.self_attn.q_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 18278400 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.3.self_attn.v_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 18343936 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.3.self_attn.v_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 18345984 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.3.self_attn.v_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 18870272 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.4.layer_norm1.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 18935808 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.4.layer_norm1.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 18937856 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.4.layer_norm2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 18939904 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.4.layer_norm2.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 18941952 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.4.mlp.fc1.bias", + "shape": [ + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8192, + "byteOffset": 18944000 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.4.mlp.fc1.q_weight", + "shape": [ + 4096, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 18952192 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.4.mlp.fc1.q_scale", + "shape": [ + 4096, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 21049344 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.4.mlp.fc2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 21311488 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.4.mlp.fc2.q_weight", + "shape": [ + 1024, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 21313536 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.4.mlp.fc2.q_scale", + "shape": [ + 1024, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 23410688 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.4.self_attn.k_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 23672832 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.4.self_attn.k_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 23674880 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.4.self_attn.k_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 24199168 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.4.self_attn.out_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 24264704 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.4.self_attn.out_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 24266752 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.4.self_attn.out_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 24791040 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.4.self_attn.q_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 24856576 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.4.self_attn.q_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 24858624 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.4.self_attn.q_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 25382912 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.4.self_attn.v_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 25448448 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.4.self_attn.v_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 25450496 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.4.self_attn.v_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 25974784 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.5.layer_norm1.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 26040320 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.5.layer_norm1.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 26042368 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.5.layer_norm2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 26044416 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.5.layer_norm2.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 26046464 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.5.mlp.fc1.bias", + "shape": [ + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8192, + "byteOffset": 26048512 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.5.mlp.fc1.q_weight", + "shape": [ + 4096, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 26056704 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.5.mlp.fc1.q_scale", + "shape": [ + 4096, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 28153856 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.5.mlp.fc2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 28416000 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.5.mlp.fc2.q_weight", + "shape": [ + 1024, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 28418048 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.5.mlp.fc2.q_scale", + "shape": [ + 1024, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 30515200 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.5.self_attn.k_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 30777344 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.5.self_attn.k_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 30779392 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.5.self_attn.k_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 31303680 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.5.self_attn.out_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 31369216 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.5.self_attn.out_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 31371264 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.5.self_attn.out_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 31895552 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.5.self_attn.q_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 31961088 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.5.self_attn.q_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 31963136 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.5.self_attn.q_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 32487424 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.5.self_attn.v_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 32552960 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.5.self_attn.v_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 32555008 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.5.self_attn.v_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 33079296 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.6.layer_norm1.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 33144832 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.6.layer_norm1.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 33146880 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.6.layer_norm2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 33148928 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.6.layer_norm2.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 33150976 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.6.mlp.fc1.bias", + "shape": [ + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8192, + "byteOffset": 33153024 + } + ], + "md5sum": "16f33cc460696f29368f48352e15c0cb" + }, + { + "dataPath": "params_shard_86.bin", + "format": "raw-shard", + "nbytes": 28416000, + "records": [ + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.6.mlp.fc1.q_weight", + "shape": [ + 4096, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 0 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.6.mlp.fc1.q_scale", + "shape": [ + 4096, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 2097152 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.6.mlp.fc2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 2359296 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.6.mlp.fc2.q_weight", + "shape": [ + 1024, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 2361344 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.6.mlp.fc2.q_scale", + "shape": [ + 1024, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 4458496 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.6.self_attn.k_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 4720640 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.6.self_attn.k_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 4722688 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.6.self_attn.k_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 5246976 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.6.self_attn.out_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 5312512 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.6.self_attn.out_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 5314560 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.6.self_attn.out_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 5838848 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.6.self_attn.q_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 5904384 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.6.self_attn.q_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 5906432 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.6.self_attn.q_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 6430720 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.6.self_attn.v_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 6496256 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.6.self_attn.v_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 6498304 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.6.self_attn.v_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 7022592 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.7.layer_norm1.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 7088128 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.7.layer_norm1.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 7090176 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.7.layer_norm2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 7092224 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.7.layer_norm2.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 7094272 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.7.mlp.fc1.bias", + "shape": [ + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8192, + "byteOffset": 7096320 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.7.mlp.fc1.q_weight", + "shape": [ + 4096, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 7104512 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.7.mlp.fc1.q_scale", + "shape": [ + 4096, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 9201664 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.7.mlp.fc2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 9463808 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.7.mlp.fc2.q_weight", + "shape": [ + 1024, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 9465856 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.7.mlp.fc2.q_scale", + "shape": [ + 1024, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 11563008 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.7.self_attn.k_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 11825152 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.7.self_attn.k_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 11827200 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.7.self_attn.k_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 12351488 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.7.self_attn.out_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 12417024 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.7.self_attn.out_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 12419072 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.7.self_attn.out_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 12943360 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.7.self_attn.q_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 13008896 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.7.self_attn.q_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 13010944 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.7.self_attn.q_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 13535232 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.7.self_attn.v_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 13600768 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.7.self_attn.v_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 13602816 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.7.self_attn.v_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 14127104 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.8.layer_norm1.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 14192640 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.8.layer_norm1.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 14194688 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.8.layer_norm2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 14196736 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.8.layer_norm2.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 14198784 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.8.mlp.fc1.bias", + "shape": [ + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8192, + "byteOffset": 14200832 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.8.mlp.fc1.q_weight", + "shape": [ + 4096, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 14209024 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.8.mlp.fc1.q_scale", + "shape": [ + 4096, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 16306176 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.8.mlp.fc2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 16568320 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.8.mlp.fc2.q_weight", + "shape": [ + 1024, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 16570368 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.8.mlp.fc2.q_scale", + "shape": [ + 1024, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 18667520 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.8.self_attn.k_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 18929664 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.8.self_attn.k_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 18931712 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.8.self_attn.k_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 19456000 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.8.self_attn.out_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 19521536 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.8.self_attn.out_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 19523584 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.8.self_attn.out_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 20047872 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.8.self_attn.q_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 20113408 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.8.self_attn.q_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 20115456 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.8.self_attn.q_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 20639744 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.8.self_attn.v_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 20705280 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.8.self_attn.v_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 20707328 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.8.self_attn.v_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 21231616 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.9.layer_norm1.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 21297152 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.9.layer_norm1.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 21299200 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.9.layer_norm2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 21301248 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.9.layer_norm2.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 21303296 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.9.mlp.fc1.bias", + "shape": [ + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8192, + "byteOffset": 21305344 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.9.mlp.fc1.q_weight", + "shape": [ + 4096, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 21313536 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.9.mlp.fc1.q_scale", + "shape": [ + 4096, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 23410688 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.9.mlp.fc2.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 23672832 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.9.mlp.fc2.q_weight", + "shape": [ + 1024, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2097152, + "byteOffset": 23674880 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.9.mlp.fc2.q_scale", + "shape": [ + 1024, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 262144, + "byteOffset": 25772032 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.9.self_attn.k_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 26034176 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.9.self_attn.k_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 26036224 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.9.self_attn.k_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 26560512 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.9.self_attn.out_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 26626048 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.9.self_attn.out_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 26628096 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.9.self_attn.out_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 27152384 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.9.self_attn.q_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 27217920 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.9.self_attn.q_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 27219968 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.9.self_attn.q_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 27744256 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.9.self_attn.v_proj.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 27809792 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.9.self_attn.v_proj.q_weight", + "shape": [ + 1024, + 128 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 524288, + "byteOffset": 27811840 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.encoder.layers.9.self_attn.v_proj.q_scale", + "shape": [ + 1024, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 65536, + "byteOffset": 28336128 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.post_layernorm.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 28401664 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.post_layernorm.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 28403712 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.pre_layrnorm.bias", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 28405760 + }, + { + "name": "vision_embed_tokens.img_processor.vision_model.pre_layrnorm.weight", + "shape": [ + 1024 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2048, + "byteOffset": 28407808 + }, + { + "name": "vision_embed_tokens.img_projection.linear_1.bias", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 28409856 + } + ], + "md5sum": "027e3a724b350d7591cb6ed3fc9ec2a0" + }, + { + "dataPath": "params_shard_87.bin", + "format": "raw-shard", + "nbytes": 12400640, + "records": [ + { + "name": "vision_embed_tokens.img_projection.linear_1.q_weight", + "shape": [ + 3072, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6291456, + "byteOffset": 0 + }, + { + "name": "vision_embed_tokens.img_projection.linear_1.q_scale", + "shape": [ + 3072, + 128 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 786432, + "byteOffset": 6291456 + }, + { + "name": "vision_embed_tokens.img_projection.linear_2.bias", + "shape": [ + 3072 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6144, + "byteOffset": 7077888 + }, + { + "name": "vision_embed_tokens.img_projection.linear_2.q_weight", + "shape": [ + 3072, + 384 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 7084032 + }, + { + "name": "vision_embed_tokens.img_projection.linear_2.q_scale", + "shape": [ + 3072, + 96 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 589824, + "byteOffset": 11802624 + }, + { + "name": "vision_embed_tokens.sub_GN", + "shape": [ + 1, + 1, + 1, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8192, + "byteOffset": 12392448 + } + ], + "md5sum": "3ea1e6cf6a4c4b7f0a2f46b23a3d754d" + } + ] +} \ No newline at end of file