LucaGroup commited on
Commit
e89d716
·
verified ·
1 Parent(s): 8eeb997

Update weights and modeling code to latest version

Browse files
config.json CHANGED
@@ -23,6 +23,7 @@
23
  "classifier_pooling_type": "value_attention",
24
  "classifier_pos_weight": 1.0,
25
  "classifier_weight": null,
 
26
  "embed_scale": 1.0,
27
  "eos_token_id": 3,
28
  "ffn_dim": 10240,
@@ -45,8 +46,7 @@
45
  "task_type": "embedding",
46
  "tie_word_embeddings": false,
47
  "token_dropout": false,
48
- "torch_dtype": "float32",
49
- "transformers_version": "4.41.2",
50
  "type_vocab_size": 2,
51
  "unk_token_id": 1,
52
  "use_embed_layer_norm": false,
 
23
  "classifier_pooling_type": "value_attention",
24
  "classifier_pos_weight": 1.0,
25
  "classifier_weight": null,
26
+ "dtype": "float32",
27
  "embed_scale": 1.0,
28
  "eos_token_id": 3,
29
  "ffn_dim": 10240,
 
46
  "task_type": "embedding",
47
  "tie_word_embeddings": false,
48
  "token_dropout": false,
49
+ "transformers_version": "4.57.0",
 
50
  "type_vocab_size": 2,
51
  "unk_token_id": 1,
52
  "use_embed_layer_norm": false,
model.safetensors.index.json CHANGED
@@ -1,5 +1,6 @@
1
  {
2
  "metadata": {
 
3
  "total_size": 6321205916
4
  },
5
  "weight_map": {
 
1
  {
2
  "metadata": {
3
+ "total_parameters": 1580300839,
4
  "total_size": 6321205916
5
  },
6
  "weight_map": {
tokenization_lucaone.py CHANGED
@@ -419,9 +419,8 @@ class LucaGPLMTokenizer(PreTrainedTokenizer):
419
  """
420
  Load tokenizer from pretrained model path (standard HuggingFace interface)
421
  """
422
- print(f"pretrained_model_name_or_path: {pretrained_model_name_or_path}")
423
  vocab_file = os.path.join(pretrained_model_name_or_path, "vocab.json")
424
- if not os.path.exists(vocab_file):
425
  print("Load from saved vocabulary (not implemented yet, use default)")
426
  return cls(vocab_type="gene_prot", **kwargs)
427
  else:
 
419
  """
420
  Load tokenizer from pretrained model path (standard HuggingFace interface)
421
  """
 
422
  vocab_file = os.path.join(pretrained_model_name_or_path, "vocab.json")
423
+ if os.path.exists(vocab_file):
424
  print("Load from saved vocabulary (not implemented yet, use default)")
425
  return cls(vocab_type="gene_prot", **kwargs)
426
  else:
tokenizer_config.json CHANGED
@@ -41,8 +41,9 @@
41
  "special": true
42
  }
43
  },
44
- "clean_up_tokenization_spaces": true,
45
  "cls_token": "[CLS]",
 
46
  "mask_token": "[MASK]",
47
  "model_max_length": 1000000000000000019884624838656,
48
  "pad_token": "[PAD]",
 
41
  "special": true
42
  }
43
  },
44
+ "clean_up_tokenization_spaces": false,
45
  "cls_token": "[CLS]",
46
+ "extra_special_tokens": {},
47
  "mask_token": "[MASK]",
48
  "model_max_length": 1000000000000000019884624838656,
49
  "pad_token": "[PAD]",