pglo commited on
Commit
8e99dc4
1 Parent(s): ff313f8

Upload config.json with huggingface_hub

Browse files
Files changed (1) hide show
  1. config.json +37 -0
config.json ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "ZambaForCausalLM"
4
+ ],
5
+ "attention_dropout": 0.0,
6
+ "attn_layer_offset": 4,
7
+ "attn_layer_period": 6,
8
+ "bos_token_id": 1,
9
+ "eos_token_id": 2,
10
+ "hidden_act": "gelu",
11
+ "hidden_mamba_act": "silu",
12
+ "hidden_size": 3712,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 14848,
15
+ "mamba_conv_bias": true,
16
+ "mamba_d_conv": 4,
17
+ "mamba_d_state": 16,
18
+ "mamba_dt_rank": 232,
19
+ "mamba_expand": 2,
20
+ "mamba_proj_bias": false,
21
+ "max_position_embeddings": 4096,
22
+ "model_type": "zamba",
23
+ "n_mamba_heads": 2,
24
+ "num_attention_heads": 16,
25
+ "num_hidden_layers": 76,
26
+ "num_key_value_heads": 16,
27
+ "num_logits_to_keep": 1,
28
+ "rms_norm_eps": 1e-05,
29
+ "rope_theta": 10000,
30
+ "sliding_window": null,
31
+ "torch_dtype": "bfloat16",
32
+ "transformers_version": "4.42.0.dev0",
33
+ "unk_token_id": 0,
34
+ "use_cache": true,
35
+ "use_mamba_kernels": true,
36
+ "vocab_size": 32000
37
+ }