arnocandel commited on
Commit
74f4d6d
1 Parent(s): b6d8495
Files changed (1) hide show
  1. config.json +4 -3
config.json CHANGED
@@ -8,7 +8,7 @@
8
  "hidden_size": 4096,
9
  "initializer_range": 0.02,
10
  "intermediate_size": 11008,
11
- "max_position_embeddings": 2048,
12
  "model_type": "llama",
13
  "num_attention_heads": 32,
14
  "num_hidden_layers": 32,
@@ -16,9 +16,10 @@
16
  "pretraining_tp": 1,
17
  "rms_norm_eps": 1e-05,
18
  "rope_scaling": null,
 
19
  "tie_word_embeddings": false,
20
- "torch_dtype": "float16",
21
- "transformers_version": "4.32.0",
22
  "use_cache": true,
23
  "vocab_size": 32016
24
  }
 
8
  "hidden_size": 4096,
9
  "initializer_range": 0.02,
10
  "intermediate_size": 11008,
11
+ "max_position_embeddings": 16384,
12
  "model_type": "llama",
13
  "num_attention_heads": 32,
14
  "num_hidden_layers": 32,
 
16
  "pretraining_tp": 1,
17
  "rms_norm_eps": 1e-05,
18
  "rope_scaling": null,
19
+ "rope_theta": 1000000,
20
  "tie_word_embeddings": false,
21
+ "torch_dtype": "bfloat16",
22
+ "transformers_version": "4.32.0.dev0",
23
  "use_cache": true,
24
  "vocab_size": 32016
25
  }