Hugging Face's logo

KoichiYasuoka
/
deberta-base-ainu-ud-goeswith

Token Classification
Transformers
PyTorch
Ainu (Japan)
deberta-v2
ainu
pos
dependency-parsing
Inference Endpoints
Model card Files Community
KoichiYasuoka commited on
Commit
0fc396c
•
1 Parent(s): 566045b

model improved

Browse files
Files changed (3) hide show
  1. config.json +52 -50
  2. maker.py +5 -4
  3. pytorch_model.bin +2 -2
config.json CHANGED
@@ -129,31 +129,32 @@
129
  "112": "SCONJ|\u63a5\u7d9a\u52a9\u8a5e|cc",
130
  "113": "SCONJ|\u63a5\u7d9a\u52a9\u8a5e|mark",
131
  "114": "SCONJ|\u63a5\u7d9a\u52a9\u8a5e|root",
132
- "115": "SCONJ|\u63a5\u7d9a\u8a5e|case",
133
- "116": "SCONJ|\u63a5\u7d9a\u8a5e|mark",
134
- "117": "SYM|_|conj",
135
- "118": "SYM|_|obj",
136
- "119": "VERB|\u4ed6\u52d5\u8a5e|acl",
137
- "120": "VERB|\u4ed6\u52d5\u8a5e|advcl",
138
- "121": "VERB|\u4ed6\u52d5\u8a5e|amod",
139
- "122": "VERB|\u4ed6\u52d5\u8a5e|ccomp",
140
- "123": "VERB|\u4ed6\u52d5\u8a5e|conj",
141
- "124": "VERB|\u4ed6\u52d5\u8a5e|parataxis",
142
- "125": "VERB|\u4ed6\u52d5\u8a5e|root",
143
- "126": "VERB|\u5b8c\u5168\u52d5\u8a5e|acl",
144
- "127": "VERB|\u5b8c\u5168\u52d5\u8a5e|advcl",
145
- "128": "VERB|\u5b8c\u5168\u52d5\u8a5e|parataxis",
146
- "129": "VERB|\u5b8c\u5168\u52d5\u8a5e|root",
147
- "130": "VERB|\u81ea\u52d5\u8a5e|acl",
148
- "131": "VERB|\u81ea\u52d5\u8a5e|advcl",
149
- "132": "VERB|\u81ea\u52d5\u8a5e|amod",
150
- "133": "VERB|\u81ea\u52d5\u8a5e|ccomp",
151
- "134": "VERB|\u81ea\u52d5\u8a5e|conj",
152
- "135": "VERB|\u81ea\u52d5\u8a5e|parataxis",
153
- "136": "VERB|\u81ea\u52d5\u8a5e|root",
154
- "137": "X|_|conj",
155
- "138": "X|_|goeswith",
156
- "139": "X|_|nsubj"
 
157
  },
158
  "initializer_range": 0.02,
159
  "intermediate_size": 3072,
@@ -273,31 +274,32 @@
273
  "SCONJ|\u63a5\u7d9a\u52a9\u8a5e|cc": 112,
274
  "SCONJ|\u63a5\u7d9a\u52a9\u8a5e|mark": 113,
275
  "SCONJ|\u63a5\u7d9a\u52a9\u8a5e|root": 114,
276
- "SCONJ|\u63a5\u7d9a\u8a5e|case": 115,
277
- "SCONJ|\u63a5\u7d9a\u8a5e|mark": 116,
278
- "SYM|_|conj": 117,
279
- "SYM|_|obj": 118,
280
- "VERB|\u4ed6\u52d5\u8a5e|acl": 119,
281
- "VERB|\u4ed6\u52d5\u8a5e|advcl": 120,
282
- "VERB|\u4ed6\u52d5\u8a5e|amod": 121,
283
- "VERB|\u4ed6\u52d5\u8a5e|ccomp": 122,
284
- "VERB|\u4ed6\u52d5\u8a5e|conj": 123,
285
- "VERB|\u4ed6\u52d5\u8a5e|parataxis": 124,
286
- "VERB|\u4ed6\u52d5\u8a5e|root": 125,
287
- "VERB|\u5b8c\u5168\u52d5\u8a5e|acl": 126,
288
- "VERB|\u5b8c\u5168\u52d5\u8a5e|advcl": 127,
289
- "VERB|\u5b8c\u5168\u52d5\u8a5e|parataxis": 128,
290
- "VERB|\u5b8c\u5168\u52d5\u8a5e|root": 129,
291
- "VERB|\u81ea\u52d5\u8a5e|acl": 130,
292
- "VERB|\u81ea\u52d5\u8a5e|advcl": 131,
293
- "VERB|\u81ea\u52d5\u8a5e|amod": 132,
294
- "VERB|\u81ea\u52d5\u8a5e|ccomp": 133,
295
- "VERB|\u81ea\u52d5\u8a5e|conj": 134,
296
- "VERB|\u81ea\u52d5\u8a5e|parataxis": 135,
297
- "VERB|\u81ea\u52d5\u8a5e|root": 136,
298
- "X|_|conj": 137,
299
- "X|_|goeswith": 138,
300
- "X|_|nsubj": 139
 
301
  },
302
  "layer_norm_eps": 1e-07,
303
  "max_position_embeddings": 512,
129
  "112": "SCONJ|\u63a5\u7d9a\u52a9\u8a5e|cc",
130
  "113": "SCONJ|\u63a5\u7d9a\u52a9\u8a5e|mark",
131
  "114": "SCONJ|\u63a5\u7d9a\u52a9\u8a5e|root",
132
+ "115": "SCONJ|\u63a5\u7d9a\u8a5e|advmod",
133
+ "116": "SCONJ|\u63a5\u7d9a\u8a5e|case",
134
+ "117": "SCONJ|\u63a5\u7d9a\u8a5e|mark",
135
+ "118": "SYM|_|conj",
136
+ "119": "SYM|_|obj",
137
+ "120": "VERB|\u4ed6\u52d5\u8a5e|acl",
138
+ "121": "VERB|\u4ed6\u52d5\u8a5e|advcl",
139
+ "122": "VERB|\u4ed6\u52d5\u8a5e|amod",
140
+ "123": "VERB|\u4ed6\u52d5\u8a5e|ccomp",
141
+ "124": "VERB|\u4ed6\u52d5\u8a5e|conj",
142
+ "125": "VERB|\u4ed6\u52d5\u8a5e|parataxis",
143
+ "126": "VERB|\u4ed6\u52d5\u8a5e|root",
144
+ "127": "VERB|\u5b8c\u5168\u52d5\u8a5e|acl",
145
+ "128": "VERB|\u5b8c\u5168\u52d5\u8a5e|advcl",
146
+ "129": "VERB|\u5b8c\u5168\u52d5\u8a5e|parataxis",
147
+ "130": "VERB|\u5b8c\u5168\u52d5\u8a5e|root",
148
+ "131": "VERB|\u81ea\u52d5\u8a5e|acl",
149
+ "132": "VERB|\u81ea\u52d5\u8a5e|advcl",
150
+ "133": "VERB|\u81ea\u52d5\u8a5e|amod",
151
+ "134": "VERB|\u81ea\u52d5\u8a5e|ccomp",
152
+ "135": "VERB|\u81ea\u52d5\u8a5e|conj",
153
+ "136": "VERB|\u81ea\u52d5\u8a5e|parataxis",
154
+ "137": "VERB|\u81ea\u52d5\u8a5e|root",
155
+ "138": "X|_|conj",
156
+ "139": "X|_|goeswith",
157
+ "140": "X|_|nsubj"
158
  },
159
  "initializer_range": 0.02,
160
  "intermediate_size": 3072,
274
  "SCONJ|\u63a5\u7d9a\u52a9\u8a5e|cc": 112,
275
  "SCONJ|\u63a5\u7d9a\u52a9\u8a5e|mark": 113,
276
  "SCONJ|\u63a5\u7d9a\u52a9\u8a5e|root": 114,
277
+ "SCONJ|\u63a5\u7d9a\u8a5e|advmod": 115,
278
+ "SCONJ|\u63a5\u7d9a\u8a5e|case": 116,
279
+ "SCONJ|\u63a5\u7d9a\u8a5e|mark": 117,
280
+ "SYM|_|conj": 118,
281
+ "SYM|_|obj": 119,
282
+ "VERB|\u4ed6\u52d5\u8a5e|acl": 120,
283
+ "VERB|\u4ed6\u52d5\u8a5e|advcl": 121,
284
+ "VERB|\u4ed6\u52d5\u8a5e|amod": 122,
285
+ "VERB|\u4ed6\u52d5\u8a5e|ccomp": 123,
286
+ "VERB|\u4ed6\u52d5\u8a5e|conj": 124,
287
+ "VERB|\u4ed6\u52d5\u8a5e|parataxis": 125,
288
+ "VERB|\u4ed6\u52d5\u8a5e|root": 126,
289
+ "VERB|\u5b8c\u5168\u52d5\u8a5e|acl": 127,
290
+ "VERB|\u5b8c\u5168\u52d5\u8a5e|advcl": 128,
291
+ "VERB|\u5b8c\u5168\u52d5\u8a5e|parataxis": 129,
292
+ "VERB|\u5b8c\u5168\u52d5\u8a5e|root": 130,
293
+ "VERB|\u81ea\u52d5\u8a5e|acl": 131,
294
+ "VERB|\u81ea\u52d5\u8a5e|advcl": 132,
295
+ "VERB|\u81ea\u52d5\u8a5e|amod": 133,
296
+ "VERB|\u81ea\u52d5\u8a5e|ccomp": 134,
297
+ "VERB|\u81ea\u52d5\u8a5e|conj": 135,
298
+ "VERB|\u81ea\u52d5\u8a5e|parataxis": 136,
299
+ "VERB|\u81ea\u52d5\u8a5e|root": 137,
300
+ "X|_|conj": 138,
301
+ "X|_|goeswith": 139,
302
+ "X|_|nsubj": 140
303
  },
304
  "layer_norm_eps": 1e-07,
305
  "max_position_embeddings": 512,
maker.py CHANGED
@@ -15,8 +15,9 @@ class UDgoeswithDataset(object):
15
  dep,c="-|_|dep",[]
16
  for s in r:
17
  t=s.split("\t")
18
- if len(t)==10 and t[0].isdecimal():
19
- c.append(t)
 
20
  elif c!=[]:
21
  for x in [1,2]:
22
  d=list(c)
@@ -42,8 +43,8 @@ from transformers import AutoTokenizer,AutoConfig,AutoModelForTokenClassificatio
42
  tkz=AutoTokenizer.from_pretrained(src)
43
  trainDS=UDgoeswithDataset("train.conllu",tkz)
44
  lid=trainDS.label2id
45
- cfg=AutoConfig.from_pretrained(src,num_labels=len(lid),label2id=lid,id2label={i:l for l,i in lid.items()},ignore_mismatched_sizes=True)
46
- arg=TrainingArguments(num_train_epochs=3,per_device_train_batch_size=64,output_dir="/tmp",overwrite_output_dir=True,save_total_limit=2,learning_rate=5e-05,warmup_ratio=0.1)
47
  trn=Trainer(args=arg,data_collator=DataCollatorForTokenClassification(tkz),model=AutoModelForTokenClassification.from_pretrained(src,config=cfg,ignore_mismatched_sizes=True),train_dataset=trainDS)
48
  trn.train()
49
  trn.save_model(tgt)
15
  dep,c="-|_|dep",[]
16
  for s in r:
17
  t=s.split("\t")
18
+ if len(t)==10:
19
+ if t[0].isdecimal():
20
+ c.append(t)
21
  elif c!=[]:
22
  for x in [1,2]:
23
  d=list(c)
43
  tkz=AutoTokenizer.from_pretrained(src)
44
  trainDS=UDgoeswithDataset("train.conllu",tkz)
45
  lid=trainDS.label2id
46
+ cfg=AutoConfig.from_pretrained(src,num_labels=len(lid),label2id=lid,id2label={i:l for l,i in lid.items()},ignore_mismatched_sizes=True,task_specific_params=None)
47
+ arg=TrainingArguments(num_train_epochs=3,per_device_train_batch_size=16,output_dir="/tmp",overwrite_output_dir=True,save_total_limit=2,learning_rate=5e-05,warmup_ratio=0.1)
48
  trn=Trainer(args=arg,data_collator=DataCollatorForTokenClassification(tkz),model=AutoModelForTokenClassification.from_pretrained(src,config=cfg,ignore_mismatched_sizes=True),train_dataset=trainDS)
49
  trn.train()
50
  trn.save_model(tgt)
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:001a50855b2baa0d1d140f455bb1b56a1945fbd05037a16451d5464272d4a2d9
3
- size 419800339
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bcadf30041fc773bdc1d871b11411b9810f9f19c0cb52eab4350d68935f65513
3
+ size 419803411