Hugging Face's logo

pure-team
/
DeepThink-T1-Tuned

Text Generation
Chinese
slm
knowledge-distillation
dark
code
Model card Files
xet
Community
1
pure-team commited on
Commit
abaad47
·
verified ·
1 Parent(s): 8bc1c87

Update config.json

Browse files
Files changed (1) hide show
  1. config.json +20 -20
config.json CHANGED
@@ -1,21 +1,21 @@
1
- {
2
- "architectures": [
3
- "DeepSeekForCausalLM"
4
- ],
5
- "model_type": "deepseek",
6
- "vocab_size": 32000,
7
- "hidden_size": 2048,
8
- "intermediate_size": 8192,
9
- "num_hidden_layers": 36,
10
- "num_attention_heads": 16,
11
- "max_position_embeddings": 2048,
12
- "rms_norm_eps": 1e-05,
13
- "initializer_range": 0.02,
14
- "use_cache": true,
15
- "pad_token_id": 0,
16
- "bos_token_id": 1,
17
- "eos_token_id": 2,
18
- "tie_word_embeddings": false,
19
- "torch_dtype": "float32",
20
- "transformers_version": "4.36.0"
21
  }
 
1
+ {
2
+ "architectures": [
3
+ "AXI with transformers"
4
+ ],
5
+ "model_type": "slm",
6
+ "vocab_size": 32000,
7
+ "hidden_size": 2048,
8
+ "intermediate_size": 8192,
9
+ "num_hidden_layers": 36,
10
+ "num_attention_heads": 16,
11
+ "max_position_embeddings": 2048,
12
+ "rms_norm_eps": 1e-05,
13
+ "initializer_range": 0.02,
14
+ "use_cache": true,
15
+ "pad_token_id": 0,
16
+ "bos_token_id": 1,
17
+ "eos_token_id": 2,
18
+ "tie_word_embeddings": false,
19
+ "torch_dtype": "float32",
20
+ "transformers_version": "4.36.0"
21
  }