wejoncy commited on
Commit
e5de82a
1 Parent(s): 24263af

update config

Browse files
Files changed (1) hide show
  1. config.json +19 -18
config.json CHANGED
@@ -22,7 +22,21 @@
22
  "num_hidden_layers": 80,
23
  "num_key_value_heads": 8,
24
  "pretraining_tp": 1,
25
- "quant_config": {
 
 
 
 
 
 
 
 
 
 
 
 
 
 
26
  "model.layers.0.mlp.down_proj": {
27
  "bias": null,
28
  "enable_norm": true,
@@ -13462,20 +13476,7 @@
13462
  -1,
13463
  8
13464
  ]
13465
- }
13466
- },
13467
- "rms_norm_eps": 1e-05,
13468
- "rope_scaling": {
13469
- "factor": 8.0,
13470
- "high_freq_factor": 4.0,
13471
- "low_freq_factor": 1.0,
13472
- "original_max_position_embeddings": 8192,
13473
- "rope_type": "llama3"
13474
- },
13475
- "rope_theta": 500000.0,
13476
- "tie_word_embeddings": false,
13477
- "torch_dtype": "bfloat16",
13478
- "transformers_version": "4.44.2",
13479
- "use_cache": true,
13480
- "vocab_size": 128256
13481
- }
 
22
  "num_hidden_layers": 80,
23
  "num_key_value_heads": 8,
24
  "pretraining_tp": 1,
25
+ "rms_norm_eps": 1e-05,
26
+ "rope_scaling": {
27
+ "factor": 8.0,
28
+ "high_freq_factor": 4.0,
29
+ "low_freq_factor": 1.0,
30
+ "original_max_position_embeddings": 8192,
31
+ "rope_type": "llama3"
32
+ },
33
+ "rope_theta": 500000.0,
34
+ "tie_word_embeddings": false,
35
+ "torch_dtype": "bfloat16",
36
+ "transformers_version": "4.44.2",
37
+ "use_cache": true,
38
+ "vocab_size": 128256,
39
+ "quantization_config": {
40
  "model.layers.0.mlp.down_proj": {
41
  "bias": null,
42
  "enable_norm": true,
 
13476
  -1,
13477
  8
13478
  ]
13479
+ },
13480
+ "quant_method": "vptq"
13481
+ }
13482
+ }