Text Generation
Transformers
Safetensors
zamba2
Inference Endpoints
pglo commited on
Commit
99f5f18
1 Parent(s): d828eb6

Upload folder using huggingface_hub

Browse files
config.json ADDED
@@ -0,0 +1,92 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bias_linear": false,
3
+ "architectures": [
4
+ "Zamba2ForCausalLM"
5
+ ],
6
+ "attention_dropout": 0.0,
7
+ "bias_gelu_fusion": false,
8
+ "bos_token_id": 1,
9
+ "conv_dimension": 4,
10
+ "eos_token_id": 2,
11
+ "expansion_factor": 2,
12
+ "ffn_hidden_size": 10240,
13
+ "gated_linear_unit": true,
14
+ "hidden_size": 2560,
15
+ "initializer_range": 0.02,
16
+ "kv_channels": 80,
17
+ "layers_block_type": [
18
+ "m",
19
+ "m",
20
+ "m",
21
+ "m",
22
+ "m",
23
+ "m",
24
+ "g",
25
+ "m",
26
+ "m",
27
+ "m",
28
+ "m",
29
+ "m",
30
+ "g",
31
+ "m",
32
+ "m",
33
+ "m",
34
+ "m",
35
+ "m",
36
+ "g",
37
+ "m",
38
+ "m",
39
+ "m",
40
+ "m",
41
+ "m",
42
+ "g",
43
+ "m",
44
+ "m",
45
+ "m",
46
+ "m",
47
+ "m",
48
+ "g",
49
+ "m",
50
+ "m",
51
+ "m",
52
+ "m",
53
+ "m",
54
+ "g",
55
+ "m",
56
+ "m",
57
+ "m",
58
+ "m",
59
+ "m",
60
+ "g",
61
+ "m",
62
+ "m",
63
+ "m",
64
+ "m",
65
+ "g",
66
+ "m",
67
+ "m",
68
+ "m",
69
+ "g",
70
+ "m",
71
+ "m"
72
+ ],
73
+ "lora_rank": 128,
74
+ "mamba_headdim": 64,
75
+ "model_type": "zamba2",
76
+ "num_attention_heads": 32,
77
+ "num_hidden_layers": 54,
78
+ "num_key_value_heads": 32,
79
+ "num_logits_to_keep": 1,
80
+ "num_query_groups": 32,
81
+ "pad_token_id": 0,
82
+ "rms_norm_eps": 1e-05,
83
+ "sliding_window": null,
84
+ "state_size": 64,
85
+ "torch_dtype": "bfloat16",
86
+ "transformers_version": "4.43.0.dev0",
87
+ "use_cache": true,
88
+ "use_low_rank_mamba_proj": false,
89
+ "use_mamba_kernels": true,
90
+ "use_shared_block_lora": true,
91
+ "vocab_size": 32000
92
+ }
generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "eos_token_id": 2,
5
+ "pad_token_id": 0,
6
+ "transformers_version": "4.43.0.dev0"
7
+ }
model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7366f9dabf8bbba25066b62aec2a870ec3f9bae910ee69997462453c0b9e8ec7
3
+ size 383523184