starmpcc commited on
Commit
5a3465c
·
verified ·
1 Parent(s): c0db165

Update config.json

Browse files
Files changed (1) hide show
  1. config.json +1 -18
config.json CHANGED
@@ -1,45 +1,28 @@
1
  {
2
- "alibi": false,
3
- "alibi_bias_max": 8,
4
  "architectures": [
5
  "LlamaForCausalLM"
6
  ],
7
  "attention_bias": false,
8
  "attention_dropout": 0.0,
9
  "bos_token_id": 128000,
10
- "causal_bias": false,
11
- "desired_length": 8192,
12
  "eos_token_id": 128001,
13
- "fire": false,
14
  "head_dim": 64,
15
  "hidden_act": "silu",
16
  "hidden_size": 2048,
17
  "initializer_range": 0.02,
18
- "input_layernorm": true,
19
  "intermediate_size": 5632,
20
  "max_position_embeddings": 131072,
21
  "mlp_bias": false,
22
  "model_type": "llama",
23
- "norm_type": "ln",
24
  "num_attention_heads": 32,
25
  "num_hidden_layers": 22,
26
  "num_key_value_heads": 8,
27
- "pretraining_length": 1024,
28
  "pretraining_tp": 1,
29
- "qk_norm": false,
30
  "rms_norm_eps": 1e-05,
31
- "rope": false,
32
- "rope_fraction": 1.0,
33
- "rope_mix": "head",
34
- "rope_scaling": null,
35
- "rope_theta": 500.0,
36
- "sigmoid_activation": false,
37
- "sinusoidal": false,
38
- "ssmax": false,
39
  "tie_word_embeddings": false,
40
  "torch_dtype": "bfloat16",
41
  "transformers_version": "4.52.4",
42
  "use_cache": false,
43
- "v_norm": false,
44
  "vocab_size": 128256
45
  }
 
1
  {
 
 
2
  "architectures": [
3
  "LlamaForCausalLM"
4
  ],
5
  "attention_bias": false,
6
  "attention_dropout": 0.0,
7
  "bos_token_id": 128000,
 
 
8
  "eos_token_id": 128001,
 
9
  "head_dim": 64,
10
  "hidden_act": "silu",
11
  "hidden_size": 2048,
12
  "initializer_range": 0.02,
 
13
  "intermediate_size": 5632,
14
  "max_position_embeddings": 131072,
15
  "mlp_bias": false,
16
  "model_type": "llama",
 
17
  "num_attention_heads": 32,
18
  "num_hidden_layers": 22,
19
  "num_key_value_heads": 8,
 
20
  "pretraining_tp": 1,
 
21
  "rms_norm_eps": 1e-05,
22
+ "rope_theta": 10000.0
 
 
 
 
 
 
 
23
  "tie_word_embeddings": false,
24
  "torch_dtype": "bfloat16",
25
  "transformers_version": "4.52.4",
26
  "use_cache": false,
 
27
  "vocab_size": 128256
28
  }