LeroyDyer commited on
Commit
6b064f4
1 Parent(s): 0f92870

Upload config.json

Browse files
Files changed (1) hide show
  1. config.json +2 -13
config.json CHANGED
@@ -2,7 +2,6 @@
2
  "architectures": [
3
  "MistralForCausalLM"
4
  ],
5
- "attention_dropout": 0.0,
6
  "bos_token_id": 1,
7
  "eos_token_id": 2,
8
  "hidden_act": "silu",
@@ -10,10 +9,6 @@
10
  "initializer_range": 0.02,
11
  "intermediate_size": 14336,
12
  "max_position_embeddings": 32768,
13
- "max_thoughts": 10,
14
- "merged_lm_and_talk_heads": false,
15
- "merged_lm_and_think_heads": true,
16
- "merged_talk_heads": true,
17
  "model_type": "mistral",
18
  "num_attention_heads": 32,
19
  "num_hidden_layers": 32,
@@ -23,13 +18,7 @@
23
  "sliding_window": 4096,
24
  "tie_word_embeddings": false,
25
  "torch_dtype": "bfloat16",
26
- "transformers_version": "4.37.0.dev0",
27
  "use_cache": true,
28
- "use_complex_talk_head": true,
29
- "use_complex_think_head": false,
30
- "use_concat_talk_head": true,
31
- "use_shallow_talk": false,
32
- "use_shallow_think": true,
33
- "use_weighted_talk_head": true,
34
- "vocab_size": 32002
35
  }
 
2
  "architectures": [
3
  "MistralForCausalLM"
4
  ],
 
5
  "bos_token_id": 1,
6
  "eos_token_id": 2,
7
  "hidden_act": "silu",
 
9
  "initializer_range": 0.02,
10
  "intermediate_size": 14336,
11
  "max_position_embeddings": 32768,
 
 
 
 
12
  "model_type": "mistral",
13
  "num_attention_heads": 32,
14
  "num_hidden_layers": 32,
 
18
  "sliding_window": 4096,
19
  "tie_word_embeddings": false,
20
  "torch_dtype": "bfloat16",
21
+ "transformers_version": "4.34.0.dev0",
22
  "use_cache": true,
23
+ "vocab_size": 32000
 
 
 
 
 
 
24
  }