diff --git "a/ndarray-cache.json" "b/ndarray-cache.json" new file mode 100644--- /dev/null +++ "b/ndarray-cache.json" @@ -0,0 +1,6425 @@ +{ + "metadata": { + "ParamSize": 507, + "ParamBytes": 4766153728.0, + "BitsPerParam": 4.068522985105678 + }, + "records": [ + { + "dataPath": "params_shard_0.bin", + "format": "raw-shard", + "nbytes": 458752000, + "records": [ + { + "name": "model.embed_tokens.q_weight", + "shape": [ + 256000, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 458752000, + "byteOffset": 0 + } + ], + "md5sum": "7e4576c07fce908093414ffff240dcdd" + }, + { + "dataPath": "params_shard_1.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.0.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "8a734062e073c60f505e870cef9c9e64" + }, + { + "dataPath": "params_shard_2.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.0.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "30ff03988087e1a456ded808ca5c3b4d" + }, + { + "dataPath": "params_shard_3.bin", + "format": "raw-shard", + "nbytes": 31911936, + "records": [ + { + "name": "model.embed_tokens.q_scale", + "shape": [ + 256000, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 14336000, + "byteOffset": 0 + }, + { + "name": "model.layers.0.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 14336000 + }, + { + "name": "model.layers.0.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 14343168 + }, + { + "name": "model.layers.0.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 15145984 + }, + { + "name": "model.layers.0.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 16751616 + }, + { + "name": "model.layers.0.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 16758784 + }, + { + "name": "model.layers.0.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 16765952 + }, + { + "name": "model.layers.0.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 16773120 + }, + { + "name": "model.layers.0.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 31453184 + } + ], + "md5sum": "e61687e23140c4be0eb472f8b8b46d06" + }, + { + "dataPath": "params_shard_4.bin", + "format": "raw-shard", + "nbytes": 33266688, + "records": [ + { + "name": "model.layers.0.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 0 + }, + { + "name": "model.layers.0.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 7340032 + }, + { + "name": "model.layers.1.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 7569408 + }, + { + "name": "model.layers.1.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 7576576 + } + ], + "md5sum": "e1e54aac747eab875148d4984ae3e402" + }, + { + "dataPath": "params_shard_5.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.1.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "fa9ae61c4e80101f2ac34f7c6cafb64f" + }, + { + "dataPath": "params_shard_6.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.2.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "3fe14ce3b96be382bcea844fcc79d6a9" + }, + { + "dataPath": "params_shard_7.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.2.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "b6c3759974f13f1c5810a8de93ff8cb3" + }, + { + "dataPath": "params_shard_8.bin", + "format": "raw-shard", + "nbytes": 27575296, + "records": [ + { + "name": "model.layers.1.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 0 + }, + { + "name": "model.layers.1.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 802816 + }, + { + "name": "model.layers.1.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 2408448 + }, + { + "name": "model.layers.1.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 2415616 + }, + { + "name": "model.layers.1.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 2422784 + }, + { + "name": "model.layers.1.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 2429952 + }, + { + "name": "model.layers.1.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 17110016 + }, + { + "name": "model.layers.1.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 17568768 + }, + { + "name": "model.layers.1.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 24908800 + }, + { + "name": "model.layers.2.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + }, + { + "name": "model.layers.2.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 25145344 + }, + { + "name": "model.layers.2.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 25948160 + }, + { + "name": "model.layers.2.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 27553792 + }, + { + "name": "model.layers.2.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 27560960 + }, + { + "name": "model.layers.2.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 27568128 + } + ], + "md5sum": "c61a981099af5c1c88c067339f9addb9" + }, + { + "dataPath": "params_shard_9.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.3.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "9828f6599fc633c572fc30cf2ba1fae4" + }, + { + "dataPath": "params_shard_10.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.3.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "e9606185e03eb1efead4c092df5539c0" + }, + { + "dataPath": "params_shard_11.bin", + "format": "raw-shard", + "nbytes": 25145344, + "records": [ + { + "name": "model.layers.2.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.2.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.2.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.2.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.3.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.3.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.3.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 23518208 + }, + { + "name": "model.layers.3.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25123840 + }, + { + "name": "model.layers.3.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25131008 + }, + { + "name": "model.layers.3.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + } + ], + "md5sum": "a0450307675d4457191fb8a29e54d991" + }, + { + "dataPath": "params_shard_12.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.4.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "7336c7ed20a9ee95956ce177ec9f1c5c" + }, + { + "dataPath": "params_shard_13.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.4.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "118cc5d1a6fab3c53404fe4737dadf96" + }, + { + "dataPath": "params_shard_14.bin", + "format": "raw-shard", + "nbytes": 25145344, + "records": [ + { + "name": "model.layers.3.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.3.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.3.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.3.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.4.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.4.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.4.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 23518208 + }, + { + "name": "model.layers.4.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25123840 + }, + { + "name": "model.layers.4.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25131008 + }, + { + "name": "model.layers.4.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + } + ], + "md5sum": "0be024d72cd013ccf872850173b5e4c6" + }, + { + "dataPath": "params_shard_15.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.5.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "2a6e654801f367267ffdddf2b7928b63" + }, + { + "dataPath": "params_shard_16.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.5.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "90c3a2dcabb63c0713d11e1e457ce114" + }, + { + "dataPath": "params_shard_17.bin", + "format": "raw-shard", + "nbytes": 25145344, + "records": [ + { + "name": "model.layers.4.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.4.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.4.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.4.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.5.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.5.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.5.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 23518208 + }, + { + "name": "model.layers.5.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25123840 + }, + { + "name": "model.layers.5.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25131008 + }, + { + "name": "model.layers.5.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + } + ], + "md5sum": "85fdc885a64ae92e4b2fef970550b785" + }, + { + "dataPath": "params_shard_18.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.6.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "2385463efb3e6207a8d9c1c3221f846a" + }, + { + "dataPath": "params_shard_19.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.6.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "9bf221721002f35c2d045a797243de10" + }, + { + "dataPath": "params_shard_20.bin", + "format": "raw-shard", + "nbytes": 25145344, + "records": [ + { + "name": "model.layers.5.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.5.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.5.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.5.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.6.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.6.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.6.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 23518208 + }, + { + "name": "model.layers.6.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25123840 + }, + { + "name": "model.layers.6.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25131008 + }, + { + "name": "model.layers.6.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + } + ], + "md5sum": "bb64e94e9ec18d3e3629667d418c0afc" + }, + { + "dataPath": "params_shard_21.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.7.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "236c0bd1a991714f0ef9401d4e2d4d55" + }, + { + "dataPath": "params_shard_22.bin", + "format": "raw-shard", + "nbytes": 24313856, + "records": [ + { + "name": "model.layers.6.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.6.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.6.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.6.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.7.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 22708224 + } + ], + "md5sum": "82d6d0465fa8507f12a596124bb9c102" + }, + { + "dataPath": "params_shard_23.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.10.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "b0aad1cbc9f37ef38df8d96949788d2c" + }, + { + "dataPath": "params_shard_24.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.10.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "164e76e7d19882e05fdde875f297dc22" + }, + { + "dataPath": "params_shard_25.bin", + "format": "raw-shard", + "nbytes": 25145344, + "records": [ + { + "name": "model.layers.7.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.7.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.7.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.7.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.10.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.10.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.10.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 23518208 + }, + { + "name": "model.layers.10.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25123840 + }, + { + "name": "model.layers.10.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25131008 + }, + { + "name": "model.layers.10.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + } + ], + "md5sum": "48116ee57504e6aac7abe8e69aab5f6c" + }, + { + "dataPath": "params_shard_26.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.11.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "471e68b3e7adaa36dafa06e8818a4c99" + }, + { + "dataPath": "params_shard_27.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.11.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "9b05f047d0ddc5750ebc89742ada351c" + }, + { + "dataPath": "params_shard_28.bin", + "format": "raw-shard", + "nbytes": 25145344, + "records": [ + { + "name": "model.layers.10.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.10.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.10.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.10.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.11.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.11.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.11.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 23518208 + }, + { + "name": "model.layers.11.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25123840 + }, + { + "name": "model.layers.11.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25131008 + }, + { + "name": "model.layers.11.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + } + ], + "md5sum": "7fab3bc28f09cef709b61029cf8cc676" + }, + { + "dataPath": "params_shard_29.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.12.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "fb4acb9b8f0a9ccfdfb9db758eb4b4a2" + }, + { + "dataPath": "params_shard_30.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.12.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "c950b21690cc0df1fd0b387a5d53f070" + }, + { + "dataPath": "params_shard_31.bin", + "format": "raw-shard", + "nbytes": 25145344, + "records": [ + { + "name": "model.layers.11.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.11.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.11.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.11.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.12.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.12.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.12.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 23518208 + }, + { + "name": "model.layers.12.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25123840 + }, + { + "name": "model.layers.12.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25131008 + }, + { + "name": "model.layers.12.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + } + ], + "md5sum": "b91f7e72a2926aed6ec8bd246fef8b51" + }, + { + "dataPath": "params_shard_32.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.13.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "c58d01a4f9930922ed1e0777e7bf9b89" + }, + { + "dataPath": "params_shard_33.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.13.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "de8c77d182370deb00bc978f3b4c8f36" + }, + { + "dataPath": "params_shard_34.bin", + "format": "raw-shard", + "nbytes": 25145344, + "records": [ + { + "name": "model.layers.12.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.12.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.12.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.12.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.13.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.13.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.13.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 23518208 + }, + { + "name": "model.layers.13.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25123840 + }, + { + "name": "model.layers.13.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25131008 + }, + { + "name": "model.layers.13.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + } + ], + "md5sum": "f78a3d895982b4c1fc6adb564cfa1933" + }, + { + "dataPath": "params_shard_35.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.14.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "6f9880fcd353f146b53996d90a7947fa" + }, + { + "dataPath": "params_shard_36.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.14.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "4d0e5df2df24f327c0dffceb896f9cdc" + }, + { + "dataPath": "params_shard_37.bin", + "format": "raw-shard", + "nbytes": 25145344, + "records": [ + { + "name": "model.layers.13.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.13.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.13.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.13.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.14.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.14.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.14.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 23518208 + }, + { + "name": "model.layers.14.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25123840 + }, + { + "name": "model.layers.14.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25131008 + }, + { + "name": "model.layers.14.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + } + ], + "md5sum": "23172be491de27f065dbd3e1f282078e" + }, + { + "dataPath": "params_shard_38.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.15.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "f52d521b8e866d691388eaf3ff9a678f" + }, + { + "dataPath": "params_shard_39.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.15.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "222f3c2eb37551edc7bd9fb2e38153db" + }, + { + "dataPath": "params_shard_40.bin", + "format": "raw-shard", + "nbytes": 25145344, + "records": [ + { + "name": "model.layers.14.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.14.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.14.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.14.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.15.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.15.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.15.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 23518208 + }, + { + "name": "model.layers.15.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25123840 + }, + { + "name": "model.layers.15.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25131008 + }, + { + "name": "model.layers.15.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + } + ], + "md5sum": "5ea2d30fb237d6c60f500e7f2ffbf2fe" + }, + { + "dataPath": "params_shard_41.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.16.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "d239076b99037e9aaf123264d5942da0" + }, + { + "dataPath": "params_shard_42.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.16.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "ae8cb7778a87321fa8c463238933aff4" + }, + { + "dataPath": "params_shard_43.bin", + "format": "raw-shard", + "nbytes": 25145344, + "records": [ + { + "name": "model.layers.15.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.15.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.15.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.15.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.16.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.16.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.16.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 23518208 + }, + { + "name": "model.layers.16.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25123840 + }, + { + "name": "model.layers.16.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25131008 + }, + { + "name": "model.layers.16.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + } + ], + "md5sum": "5b08ec540a7c858aee330877683ea64a" + }, + { + "dataPath": "params_shard_44.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.17.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "d6e08e56ca22f2f18a600c03eb3ed1c8" + }, + { + "dataPath": "params_shard_45.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.17.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "7bd707824d4db8ea33251e57e345cb1c" + }, + { + "dataPath": "params_shard_46.bin", + "format": "raw-shard", + "nbytes": 25145344, + "records": [ + { + "name": "model.layers.16.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.16.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.16.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.16.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.17.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.17.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.17.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 23518208 + }, + { + "name": "model.layers.17.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25123840 + }, + { + "name": "model.layers.17.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25131008 + }, + { + "name": "model.layers.17.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + } + ], + "md5sum": "2e27724f5d53ab5b5bbce2a8d63289b6" + }, + { + "dataPath": "params_shard_47.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.18.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "3c4dad433112a5d0ba973ed359a4e929" + }, + { + "dataPath": "params_shard_48.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.18.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "dc6b386770b838444f42968492b5b07c" + }, + { + "dataPath": "params_shard_49.bin", + "format": "raw-shard", + "nbytes": 25145344, + "records": [ + { + "name": "model.layers.17.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.17.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.17.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.17.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.18.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.18.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.18.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 23518208 + }, + { + "name": "model.layers.18.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25123840 + }, + { + "name": "model.layers.18.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25131008 + }, + { + "name": "model.layers.18.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + } + ], + "md5sum": "18620e716961655b3eec80e475107559" + }, + { + "dataPath": "params_shard_50.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.19.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "f2ce1f6fa604f49bead2b24cff7029c9" + }, + { + "dataPath": "params_shard_51.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.19.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "ce76c88eeee3cefca1869fe638e802a3" + }, + { + "dataPath": "params_shard_52.bin", + "format": "raw-shard", + "nbytes": 25145344, + "records": [ + { + "name": "model.layers.18.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.18.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.18.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.18.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.19.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.19.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.19.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 23518208 + }, + { + "name": "model.layers.19.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25123840 + }, + { + "name": "model.layers.19.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25131008 + }, + { + "name": "model.layers.19.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + } + ], + "md5sum": "112bb1b4b7085092dbcb34764492afb2" + }, + { + "dataPath": "params_shard_53.bin", + "format": "raw-shard", + "nbytes": 22708224, + "records": [ + { + "name": "model.layers.19.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.19.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.19.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.19.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + } + ], + "md5sum": "3f4bf5622b9ea9970c7fdc45d6e0cad4" + }, + { + "dataPath": "params_shard_54.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.7.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "ce248ae228d43dcce7b9337f96c6dc92" + }, + { + "dataPath": "params_shard_55.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.8.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "8a60dec7fb94b0ca9c12da6afb00d98a" + }, + { + "dataPath": "params_shard_56.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.8.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "f10d03cd1899cf74614a73bd38b892dc" + }, + { + "dataPath": "params_shard_57.bin", + "format": "raw-shard", + "nbytes": 33546240, + "records": [ + { + "name": "model.layers.20.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.20.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.7.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 15138816 + }, + { + "name": "model.layers.7.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 15145984 + }, + { + "name": "model.layers.7.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 15948800 + }, + { + "name": "model.layers.7.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 15955968 + }, + { + "name": "model.layers.7.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 15963136 + }, + { + "name": "model.layers.8.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 15970304 + }, + { + "name": "model.layers.8.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 15977472 + }, + { + "name": "model.layers.8.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 16780288 + }, + { + "name": "model.layers.8.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 18385920 + }, + { + "name": "model.layers.8.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 18393088 + }, + { + "name": "model.layers.8.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 18400256 + }, + { + "name": "model.layers.8.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 18407424 + }, + { + "name": "model.layers.8.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 33087488 + } + ], + "md5sum": "307526fa6a0cd170423fc5e0b31ec104" + }, + { + "dataPath": "params_shard_58.bin", + "format": "raw-shard", + "nbytes": 33266688, + "records": [ + { + "name": "model.layers.8.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 0 + }, + { + "name": "model.layers.8.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 7340032 + }, + { + "name": "model.layers.9.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 7569408 + }, + { + "name": "model.layers.9.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 7576576 + } + ], + "md5sum": "281927647b9017bd9fa752fcba708a1d" + }, + { + "dataPath": "params_shard_59.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.9.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "0e33d6b360aef369d199b7dfb71fac2b" + }, + { + "dataPath": "params_shard_60.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.20.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "5b3eb11fc57eebdc6a2407c0447390d9" + }, + { + "dataPath": "params_shard_61.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.20.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "e09440c385bf81b124c52f817e749ce2" + }, + { + "dataPath": "params_shard_62.bin", + "format": "raw-shard", + "nbytes": 27575296, + "records": [ + { + "name": "model.layers.9.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 0 + }, + { + "name": "model.layers.9.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 802816 + }, + { + "name": "model.layers.9.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 2408448 + }, + { + "name": "model.layers.9.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 2415616 + }, + { + "name": "model.layers.9.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 2422784 + }, + { + "name": "model.layers.9.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 2429952 + }, + { + "name": "model.layers.9.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 17110016 + }, + { + "name": "model.layers.9.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 17568768 + }, + { + "name": "model.layers.9.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 24908800 + }, + { + "name": "model.layers.20.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + }, + { + "name": "model.layers.20.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 25145344 + }, + { + "name": "model.layers.20.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 25948160 + }, + { + "name": "model.layers.20.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 27553792 + }, + { + "name": "model.layers.20.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 27560960 + }, + { + "name": "model.layers.20.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 27568128 + } + ], + "md5sum": "558e33f3101ea1d125d466c49c182ed7" + }, + { + "dataPath": "params_shard_63.bin", + "format": "raw-shard", + "nbytes": 33266688, + "records": [ + { + "name": "model.layers.20.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 0 + }, + { + "name": "model.layers.20.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 7340032 + }, + { + "name": "model.layers.21.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 7569408 + }, + { + "name": "model.layers.21.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 7576576 + } + ], + "md5sum": "6100b2db63bee85ccd58f01498d1f66a" + }, + { + "dataPath": "params_shard_64.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.21.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "b398540c95c843a12c8a9583f44869df" + }, + { + "dataPath": "params_shard_65.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.22.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "f38c79d6faefe4c5a8a405128727cf1a" + }, + { + "dataPath": "params_shard_66.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.22.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "3ab949aec9fcb01b3bb74c344408c3ea" + }, + { + "dataPath": "params_shard_67.bin", + "format": "raw-shard", + "nbytes": 27575296, + "records": [ + { + "name": "model.layers.21.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 0 + }, + { + "name": "model.layers.21.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 802816 + }, + { + "name": "model.layers.21.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 2408448 + }, + { + "name": "model.layers.21.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 2415616 + }, + { + "name": "model.layers.21.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 2422784 + }, + { + "name": "model.layers.21.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 2429952 + }, + { + "name": "model.layers.21.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 17110016 + }, + { + "name": "model.layers.21.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 17568768 + }, + { + "name": "model.layers.21.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 24908800 + }, + { + "name": "model.layers.22.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + }, + { + "name": "model.layers.22.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 25145344 + }, + { + "name": "model.layers.22.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 25948160 + }, + { + "name": "model.layers.22.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 27553792 + }, + { + "name": "model.layers.22.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 27560960 + }, + { + "name": "model.layers.22.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 27568128 + } + ], + "md5sum": "a9e87b93666ba9cec43b68110f2ee7b2" + }, + { + "dataPath": "params_shard_68.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.23.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "8dd7c24f724b36c6e64b7492a0712308" + }, + { + "dataPath": "params_shard_69.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.23.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "5b4564610b7879b8f4fd65f7fbbd3386" + }, + { + "dataPath": "params_shard_70.bin", + "format": "raw-shard", + "nbytes": 25145344, + "records": [ + { + "name": "model.layers.22.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.22.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.22.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.22.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.23.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.23.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.23.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 23518208 + }, + { + "name": "model.layers.23.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25123840 + }, + { + "name": "model.layers.23.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25131008 + }, + { + "name": "model.layers.23.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + } + ], + "md5sum": "bfba488afce2abb1a9bed64fee346575" + }, + { + "dataPath": "params_shard_71.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.24.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "6429aa588539a2b7ad413e30b6e710ab" + }, + { + "dataPath": "params_shard_72.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.24.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "1d915fe2c351fc37b55681d485b15651" + }, + { + "dataPath": "params_shard_73.bin", + "format": "raw-shard", + "nbytes": 25145344, + "records": [ + { + "name": "model.layers.23.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.23.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.23.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.23.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.24.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.24.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.24.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 23518208 + }, + { + "name": "model.layers.24.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25123840 + }, + { + "name": "model.layers.24.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25131008 + }, + { + "name": "model.layers.24.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + } + ], + "md5sum": "0d68dfb609c3dad2bc9bbd50ddc0289c" + }, + { + "dataPath": "params_shard_74.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.25.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "971834151df674343dc48e283e53305f" + }, + { + "dataPath": "params_shard_75.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.25.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "483df670721fb77b9af3d60f52931c87" + }, + { + "dataPath": "params_shard_76.bin", + "format": "raw-shard", + "nbytes": 25145344, + "records": [ + { + "name": "model.layers.24.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.24.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.24.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.24.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.25.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.25.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.25.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 23518208 + }, + { + "name": "model.layers.25.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25123840 + }, + { + "name": "model.layers.25.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25131008 + }, + { + "name": "model.layers.25.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + } + ], + "md5sum": "6312634840a234492e54d846b2e928ac" + }, + { + "dataPath": "params_shard_77.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.26.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "ed47fa87d985fd697c3e19d4ef74b218" + }, + { + "dataPath": "params_shard_78.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.26.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "9a081e64d6fc4dc68b49c4fc65175c73" + }, + { + "dataPath": "params_shard_79.bin", + "format": "raw-shard", + "nbytes": 25145344, + "records": [ + { + "name": "model.layers.25.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.25.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.25.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.25.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.26.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.26.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.26.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 23518208 + }, + { + "name": "model.layers.26.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25123840 + }, + { + "name": "model.layers.26.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25131008 + }, + { + "name": "model.layers.26.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + } + ], + "md5sum": "3a74dfc7aa0ff4dad17e5ae8c6315860" + }, + { + "dataPath": "params_shard_80.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.27.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "6a98a3bdbec57bd0d35faf0cede43fe6" + }, + { + "dataPath": "params_shard_81.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.27.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "287aef2603ee42e50a8d423368b394a9" + }, + { + "dataPath": "params_shard_82.bin", + "format": "raw-shard", + "nbytes": 25145344, + "records": [ + { + "name": "model.layers.26.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.26.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.26.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.26.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.27.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.27.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.27.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 23518208 + }, + { + "name": "model.layers.27.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25123840 + }, + { + "name": "model.layers.27.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25131008 + }, + { + "name": "model.layers.27.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + } + ], + "md5sum": "5d3024a58b3db130527e4a6d4dafcc0b" + }, + { + "dataPath": "params_shard_83.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.28.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "41f35587f245bf88d1450702024485d2" + }, + { + "dataPath": "params_shard_84.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.28.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "a34fb90fd72771132d38587d422e3f93" + }, + { + "dataPath": "params_shard_85.bin", + "format": "raw-shard", + "nbytes": 25145344, + "records": [ + { + "name": "model.layers.27.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.27.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.27.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.27.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.28.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.28.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.28.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 23518208 + }, + { + "name": "model.layers.28.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25123840 + }, + { + "name": "model.layers.28.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25131008 + }, + { + "name": "model.layers.28.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + } + ], + "md5sum": "e6a7678bd7d9f782dc7d99c2a281a8fe" + }, + { + "dataPath": "params_shard_86.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.29.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "84fb15a8d067ef8f797fbd92103692c2" + }, + { + "dataPath": "params_shard_87.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.29.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "6e3ae4699e6aaa61b5bde0b1641b23ff" + }, + { + "dataPath": "params_shard_88.bin", + "format": "raw-shard", + "nbytes": 25145344, + "records": [ + { + "name": "model.layers.28.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.28.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.28.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.28.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.29.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.29.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.29.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 23518208 + }, + { + "name": "model.layers.29.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25123840 + }, + { + "name": "model.layers.29.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25131008 + }, + { + "name": "model.layers.29.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + } + ], + "md5sum": "6290040a085ee489c75a7593b840ab69" + }, + { + "dataPath": "params_shard_89.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.30.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "e1fbd6a0f7ae693bdb17049dc3c83119" + }, + { + "dataPath": "params_shard_90.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.30.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "f063ac2632907471a3a74666b50d1494" + }, + { + "dataPath": "params_shard_91.bin", + "format": "raw-shard", + "nbytes": 25145344, + "records": [ + { + "name": "model.layers.29.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.29.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.29.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.29.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.30.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.30.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.30.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 23518208 + }, + { + "name": "model.layers.30.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25123840 + }, + { + "name": "model.layers.30.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25131008 + }, + { + "name": "model.layers.30.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + } + ], + "md5sum": "d84521924e23857c659bcff3ffc0f841" + }, + { + "dataPath": "params_shard_92.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.31.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "8051beed966679d643cde5f3163698d9" + }, + { + "dataPath": "params_shard_93.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.31.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "2a7ba5b9541a1b9882bf8e788dbf9fa9" + }, + { + "dataPath": "params_shard_94.bin", + "format": "raw-shard", + "nbytes": 25145344, + "records": [ + { + "name": "model.layers.30.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.30.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.30.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.30.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.31.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.31.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.31.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 23518208 + }, + { + "name": "model.layers.31.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25123840 + }, + { + "name": "model.layers.31.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25131008 + }, + { + "name": "model.layers.31.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + } + ], + "md5sum": "238e15de479ce1da5a6d01a5861dfb5e" + }, + { + "dataPath": "params_shard_95.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.32.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "a8d06b62f5bf25bd6af84c565646ae8a" + }, + { + "dataPath": "params_shard_96.bin", + "format": "raw-shard", + "nbytes": 24313856, + "records": [ + { + "name": "model.layers.31.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.31.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.31.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.31.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.32.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 22708224 + } + ], + "md5sum": "3f5b42df2f686edcae01c81c2a983319" + }, + { + "dataPath": "params_shard_97.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.32.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "d75981f9f983db23f433c80e9441df06" + }, + { + "dataPath": "params_shard_98.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.33.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "bd5928a6143314a18cdd0af8f2ff133b" + }, + { + "dataPath": "params_shard_99.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.33.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "78a2a1ea3ebb0a15d19fb54be4b3b54a" + }, + { + "dataPath": "params_shard_100.bin", + "format": "raw-shard", + "nbytes": 25976832, + "records": [ + { + "name": "model.layers.32.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.32.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.32.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.32.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.32.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.32.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.32.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 23518208 + }, + { + "name": "model.layers.32.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 23525376 + }, + { + "name": "model.layers.32.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 23532544 + }, + { + "name": "model.layers.33.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 23539712 + }, + { + "name": "model.layers.33.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 23546880 + }, + { + "name": "model.layers.33.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 24349696 + }, + { + "name": "model.layers.33.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25955328 + }, + { + "name": "model.layers.33.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25962496 + }, + { + "name": "model.layers.33.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25969664 + } + ], + "md5sum": "9533b4d3e24e30eeb9eab11411e7c457" + }, + { + "dataPath": "params_shard_101.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.34.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "b9659841c71866463467e37b004a47e6" + }, + { + "dataPath": "params_shard_102.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.34.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "912b36721a6ede6d3e52eddd7bfda820" + }, + { + "dataPath": "params_shard_103.bin", + "format": "raw-shard", + "nbytes": 25145344, + "records": [ + { + "name": "model.layers.33.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.33.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.33.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.33.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.34.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.34.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.34.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 23518208 + }, + { + "name": "model.layers.34.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25123840 + }, + { + "name": "model.layers.34.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25131008 + }, + { + "name": "model.layers.34.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + } + ], + "md5sum": "89296e94ab48388c32d91b63543300cf" + }, + { + "dataPath": "params_shard_104.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.35.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "94558a53493669990a2c22254cd3cb7e" + }, + { + "dataPath": "params_shard_105.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.35.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "08e20bd8ddb19bb9d7a15bb7d79ceeb1" + }, + { + "dataPath": "params_shard_106.bin", + "format": "raw-shard", + "nbytes": 25145344, + "records": [ + { + "name": "model.layers.34.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.34.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.34.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.34.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.35.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.35.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.35.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 23518208 + }, + { + "name": "model.layers.35.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25123840 + }, + { + "name": "model.layers.35.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25131008 + }, + { + "name": "model.layers.35.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + } + ], + "md5sum": "3ff946ebea2c14a447ad697730167238" + }, + { + "dataPath": "params_shard_107.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.36.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "95c726681e4b0a132bfa8caa298e893d" + }, + { + "dataPath": "params_shard_108.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.36.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "df819cb146b9beda4009f2a3743a30e6" + }, + { + "dataPath": "params_shard_109.bin", + "format": "raw-shard", + "nbytes": 25145344, + "records": [ + { + "name": "model.layers.35.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.35.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.35.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.35.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.36.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.36.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.36.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 23518208 + }, + { + "name": "model.layers.36.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25123840 + }, + { + "name": "model.layers.36.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25131008 + }, + { + "name": "model.layers.36.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + } + ], + "md5sum": "0aebc026d5c65e740aa5131798cd49e2" + }, + { + "dataPath": "params_shard_110.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.37.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "f50a2c3244567b148c0aecf50cf98638" + }, + { + "dataPath": "params_shard_111.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.37.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "632aa94539eefb8a7d2b7fa61b01ffb1" + }, + { + "dataPath": "params_shard_112.bin", + "format": "raw-shard", + "nbytes": 25145344, + "records": [ + { + "name": "model.layers.36.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.36.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.36.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.36.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.37.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.37.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.37.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 23518208 + }, + { + "name": "model.layers.37.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25123840 + }, + { + "name": "model.layers.37.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25131008 + }, + { + "name": "model.layers.37.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + } + ], + "md5sum": "dba6b5156dfb98aaf4570b545d28d1e5" + }, + { + "dataPath": "params_shard_113.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.38.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "bfd68cf560e6c68b08f53f10b262d396" + }, + { + "dataPath": "params_shard_114.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.38.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "656089d087a94655cd36cdd5ddfcdef1" + }, + { + "dataPath": "params_shard_115.bin", + "format": "raw-shard", + "nbytes": 25145344, + "records": [ + { + "name": "model.layers.37.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.37.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.37.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.37.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.38.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.38.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.38.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 23518208 + }, + { + "name": "model.layers.38.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25123840 + }, + { + "name": "model.layers.38.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25131008 + }, + { + "name": "model.layers.38.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + } + ], + "md5sum": "85476d61319e8be2410b8bcbb74cf89e" + }, + { + "dataPath": "params_shard_116.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.39.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "415615c4203cb7b11b2575fe2b4c6b5f" + }, + { + "dataPath": "params_shard_117.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.39.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "a9ba0f3157f19c9a3992d50a58472203" + }, + { + "dataPath": "params_shard_118.bin", + "format": "raw-shard", + "nbytes": 25145344, + "records": [ + { + "name": "model.layers.38.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.38.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.38.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.38.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.39.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.39.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.39.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 23518208 + }, + { + "name": "model.layers.39.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25123840 + }, + { + "name": "model.layers.39.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25131008 + }, + { + "name": "model.layers.39.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + } + ], + "md5sum": "1882f5f2e1a1cb37f82ede0c561f7217" + }, + { + "dataPath": "params_shard_119.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.40.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "ed1ae69e412681d303284218956721e2" + }, + { + "dataPath": "params_shard_120.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.40.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "f76e7ba87a65cf9536c19ee7cc167dc8" + }, + { + "dataPath": "params_shard_121.bin", + "format": "raw-shard", + "nbytes": 25145344, + "records": [ + { + "name": "model.layers.39.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.39.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.39.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.39.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.40.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.40.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.40.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 23518208 + }, + { + "name": "model.layers.40.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25123840 + }, + { + "name": "model.layers.40.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25131008 + }, + { + "name": "model.layers.40.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + } + ], + "md5sum": "0ef28051d6f025660e808cfd1bd55610" + }, + { + "dataPath": "params_shard_122.bin", + "format": "raw-shard", + "nbytes": 25690112, + "records": [ + { + "name": "model.layers.41.mlp.down_proj.q_weight", + "shape": [ + 3584, + 1792 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 25690112, + "byteOffset": 0 + } + ], + "md5sum": "fba921ebe23a12c235cedcab48554fe9" + }, + { + "dataPath": "params_shard_123.bin", + "format": "raw-shard", + "nbytes": 51380224, + "records": [ + { + "name": "model.layers.41.mlp.gate_up_proj.q_weight", + "shape": [ + 28672, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 51380224, + "byteOffset": 0 + } + ], + "md5sum": "44baa194024f6e33d812f78c541a18b6" + }, + { + "dataPath": "params_shard_124.bin", + "format": "raw-shard", + "nbytes": 25145344, + "records": [ + { + "name": "model.layers.40.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.40.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.40.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.40.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.layers.41.input_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + }, + { + "name": "model.layers.41.mlp.down_proj.q_scale", + "shape": [ + 3584, + 112 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 802816, + "byteOffset": 22715392 + }, + { + "name": "model.layers.41.mlp.gate_up_proj.q_scale", + "shape": [ + 28672, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1605632, + "byteOffset": 23518208 + }, + { + "name": "model.layers.41.post_attention_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25123840 + }, + { + "name": "model.layers.41.post_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25131008 + }, + { + "name": "model.layers.41.pre_feedforward_layernorm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 25138176 + } + ], + "md5sum": "276e3f6ce31b75ac3ae4276403e15afe" + }, + { + "dataPath": "params_shard_125.bin", + "format": "raw-shard", + "nbytes": 22715392, + "records": [ + { + "name": "model.layers.41.self_attn.qkv_proj.q_weight", + "shape": [ + 8192, + 448 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 14680064, + "byteOffset": 0 + }, + { + "name": "model.layers.41.self_attn.qkv_proj.q_scale", + "shape": [ + 8192, + 28 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 458752, + "byteOffset": 14680064 + }, + { + "name": "model.layers.41.self_attn.o_proj.q_weight", + "shape": [ + 3584, + 512 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7340032, + "byteOffset": 15138816 + }, + { + "name": "model.layers.41.self_attn.o_proj.q_scale", + "shape": [ + 3584, + 32 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 229376, + "byteOffset": 22478848 + }, + { + "name": "model.norm.weight", + "shape": [ + 3584 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7168, + "byteOffset": 22708224 + } + ], + "md5sum": "ca54e73e86e70c9c1ce3d5174d86c928" + } + ] +} \ No newline at end of file