SREDWise commited on
Commit
4ab86d7
1 Parent(s): f94ef68

Upload MistralForCausalLM

Browse files
Files changed (2) hide show
  1. config.json +10 -2
  2. pytorch_model.bin +2 -2
config.json CHANGED
@@ -4,7 +4,7 @@
4
  "activation_dropout": 0.0,
5
  "activation_function": "relu",
6
  "architectures": [
7
- "OPTForCausalLM"
8
  ],
9
  "attention_dropout": 0.0,
10
  "bos_token_id": 2,
@@ -13,16 +13,24 @@
13
  "enable_bias": true,
14
  "eos_token_id": 2,
15
  "ffn_dim": 3072,
 
16
  "hidden_size": 768,
17
  "init_std": 0.02,
 
 
18
  "layer_norm_elementwise_affine": true,
19
  "layerdrop": 0.0,
20
  "max_position_embeddings": 2048,
21
- "model_type": "opt",
22
  "num_attention_heads": 12,
23
  "num_hidden_layers": 12,
 
24
  "pad_token_id": 1,
25
  "prefix": "</s>",
 
 
 
 
26
  "torch_dtype": "float32",
27
  "transformers_version": "4.34.0",
28
  "use_cache": true,
 
4
  "activation_dropout": 0.0,
5
  "activation_function": "relu",
6
  "architectures": [
7
+ "MistralForCausalLM"
8
  ],
9
  "attention_dropout": 0.0,
10
  "bos_token_id": 2,
 
13
  "enable_bias": true,
14
  "eos_token_id": 2,
15
  "ffn_dim": 3072,
16
+ "hidden_act": "silu",
17
  "hidden_size": 768,
18
  "init_std": 0.02,
19
+ "initializer_range": 0.02,
20
+ "intermediate_size": 14336,
21
  "layer_norm_elementwise_affine": true,
22
  "layerdrop": 0.0,
23
  "max_position_embeddings": 2048,
24
+ "model_type": "mistral",
25
  "num_attention_heads": 12,
26
  "num_hidden_layers": 12,
27
+ "num_key_value_heads": 8,
28
  "pad_token_id": 1,
29
  "prefix": "</s>",
30
+ "rms_norm_eps": 1e-06,
31
+ "rope_theta": 10000.0,
32
+ "sliding_window": 4096,
33
+ "tie_word_embeddings": false,
34
  "torch_dtype": "float32",
35
  "transformers_version": "4.34.0",
36
  "use_cache": true,
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:d49f3e861a560dd71a0cc1a885d918a3b709033f1244402b8e394155d39f7c9b
3
- size 501021082
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cd2429156f122a87307530b1222cb15b7726440278647f72d1e8833784686b34
3
+ size 1988802630