LiqunMa commited on
Commit
740eb40
·
verified ·
1 Parent(s): 3c37a9f

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +2 -3
README.md CHANGED
@@ -26,7 +26,7 @@ We use [AmberDateset](https://huggingface.co/datasets/LLM360/AmberDatasets) to t
26
 
27
  ## Result
28
 
29
- <img src=https://huggingface.co/LiqunMa/FBI-LLM_7B/resolve/main/main_result.jpg width="70%" />
30
  <!-- ![image](https://huggingface.co/LiqunMa/FBI-LLM_7B/resolve/main/main_result.jpg =200x) -->
31
 
32
  ## How to use
@@ -53,8 +53,7 @@ def load_model(model_size, model_dir):
53
  for p in ckpt_plist:
54
  weight_dict = torch.load(p)
55
  for k,v in _weight_dict.items():
56
- if 'self_attn.rotary_emb.inv_freq' not in k:
57
- weight_dict[k] = v
58
 
59
  model.load_state_dict(weight_dict)
60
  for param in model.parameters():
 
26
 
27
  ## Result
28
 
29
+ <img src=https://huggingface.co/LiqunMa/FBI-LLM_7B/resolve/main/main_result.jpg width="90%" />
30
  <!-- ![image](https://huggingface.co/LiqunMa/FBI-LLM_7B/resolve/main/main_result.jpg =200x) -->
31
 
32
  ## How to use
 
53
  for p in ckpt_plist:
54
  weight_dict = torch.load(p)
55
  for k,v in _weight_dict.items():
56
+ weight_dict[k] = v
 
57
 
58
  model.load_state_dict(weight_dict)
59
  for param in model.parameters():