appvoid commited on
Commit
debe9db
1 Parent(s): 9e0e99a

Update config.json

Browse files
Files changed (1) hide show
  1. config.json +4 -8
config.json CHANGED
@@ -1,5 +1,5 @@
1
  {
2
- "_name_or_path": "appvoid/arco-original",
3
  "architectures": [
4
  "LlamaForCausalLM"
5
  ],
@@ -12,21 +12,17 @@
12
  "initializer_range": 0.02,
13
  "intermediate_size": 4096,
14
  "max_position_embeddings": 8192,
15
- "mlp_bias": false,
16
  "model_type": "llama",
17
  "num_attention_heads": 16,
18
  "num_hidden_layers": 16,
19
  "num_key_value_heads": 8,
20
  "pad_token_id": 0,
21
- "pretraining_tp": 1,
22
  "rms_norm_eps": 1e-05,
23
- "rope_scaling": null,
24
  "rope_theta": 100000,
25
  "sliding_window": null,
26
  "tie_word_embeddings": false,
27
- "torch_dtype": "float16",
28
- "transformers_version": "4.41.2",
29
- "unsloth_version": "2024.7",
30
  "use_cache": true,
31
  "vocab_size": 32000
32
- }
 
1
  {
2
+ "_name_or_path": "appvoid/arco",
3
  "architectures": [
4
  "LlamaForCausalLM"
5
  ],
 
12
  "initializer_range": 0.02,
13
  "intermediate_size": 4096,
14
  "max_position_embeddings": 8192,
 
15
  "model_type": "llama",
16
  "num_attention_heads": 16,
17
  "num_hidden_layers": 16,
18
  "num_key_value_heads": 8,
19
  "pad_token_id": 0,
 
20
  "rms_norm_eps": 1e-05,
 
21
  "rope_theta": 100000,
22
  "sliding_window": null,
23
  "tie_word_embeddings": false,
24
+ "torch_dtype": "bfloat16",
25
+ "transformers_version": "4.42.3",
 
26
  "use_cache": true,
27
  "vocab_size": 32000
28
+ }