Ray0323 commited on
Commit
4a3eac9
1 Parent(s): 978e6a8

Upload 2 files

Browse files
Files changed (2) hide show
  1. config.json +3 -6
  2. generation_config.json +1 -1
config.json CHANGED
@@ -1,26 +1,23 @@
1
  {
 
2
  "architectures": [
3
  "LlamaForCausalLM"
4
  ],
5
  "bos_token_id": 1,
6
  "eos_token_id": 2,
7
- "hidden_act": "swish",
8
  "hidden_size": 5120,
9
  "initializer_range": 0.02,
10
  "intermediate_size": 13824,
11
- "layer_norm_eps": 1e-05,
12
  "max_position_embeddings": 2048,
13
  "model_type": "llama",
14
  "num_attention_heads": 40,
15
  "num_hidden_layers": 40,
16
  "pad_token_id": 0,
17
  "rms_norm_eps": 1e-06,
18
- "rotary_emb_base": 10000,
19
- "rotary_pct": 1,
20
  "tie_word_embeddings": false,
21
  "torch_dtype": "float16",
22
- "transformers_version": "4.28.1",
23
  "use_cache": true,
24
- "use_parallel_residual": false,
25
  "vocab_size": 39424
26
  }
 
1
  {
2
+ "_name_or_path": "/cognitive_comp/wanghao/models/llama_sft/llama_13b_v2_s2_6600_rlhf_0515_kl_RM13BB0510_step49_hf/",
3
  "architectures": [
4
  "LlamaForCausalLM"
5
  ],
6
  "bos_token_id": 1,
7
  "eos_token_id": 2,
8
+ "hidden_act": "silu",
9
  "hidden_size": 5120,
10
  "initializer_range": 0.02,
11
  "intermediate_size": 13824,
 
12
  "max_position_embeddings": 2048,
13
  "model_type": "llama",
14
  "num_attention_heads": 40,
15
  "num_hidden_layers": 40,
16
  "pad_token_id": 0,
17
  "rms_norm_eps": 1e-06,
 
 
18
  "tie_word_embeddings": false,
19
  "torch_dtype": "float16",
20
+ "transformers_version": "4.29.0.dev0",
21
  "use_cache": true,
 
22
  "vocab_size": 39424
23
  }
generation_config.json CHANGED
@@ -3,5 +3,5 @@
3
  "bos_token_id": 1,
4
  "eos_token_id": 2,
5
  "pad_token_id": 0,
6
- "transformers_version": "4.28.1"
7
  }
 
3
  "bos_token_id": 1,
4
  "eos_token_id": 2,
5
  "pad_token_id": 0,
6
+ "transformers_version": "4.29.0.dev0"
7
  }