Hugging Face's logo

howdoiuse-keyboard
/
indian-address-parser-model

Token Classification
Transformers
PyTorch
English
Hindi
ner
address-parsing
indian-addresses
bert
crf
Eval Results (legacy)
Model card Files
xet
 Community
2
x commited on
Commit
bea79cc
·
verified ·
1 Parent(s): 877c442

Upload IndicBERTv2-CRF model (80% F1 on Indian addresses)

Browse files
Files changed (6) hide show
  1. config.json +73 -0
  2. pytorch_model.bin +3 -0
  3. test_results.json +6 -0
  4. tokenizer.json +0 -0
  5. tokenizer_config.json +11 -0
  6. training_info.json +14 -0
config.json ADDED
@@ -0,0 +1,73 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_name": "ai4bharat/IndicBERTv2-SS",
3
+ "num_labels": 31,
4
+ "use_crf": true,
5
+ "hidden_size": 768,
6
+ "classifier_dropout": 0.1,
7
+ "id2label": {
8
+ "0": "O",
9
+ "1": "B-AREA",
10
+ "2": "B-SUBAREA",
11
+ "3": "B-HOUSE_NUMBER",
12
+ "4": "B-SECTOR",
13
+ "5": "B-GALI",
14
+ "6": "B-COLONY",
15
+ "7": "B-BLOCK",
16
+ "8": "B-CAMP",
17
+ "9": "B-POLE",
18
+ "10": "B-KHASRA",
19
+ "11": "B-FLOOR",
20
+ "12": "B-PLOT",
21
+ "13": "B-PINCODE",
22
+ "14": "B-CITY",
23
+ "15": "B-STATE",
24
+ "16": "I-AREA",
25
+ "17": "I-SUBAREA",
26
+ "18": "I-HOUSE_NUMBER",
27
+ "19": "I-SECTOR",
28
+ "20": "I-GALI",
29
+ "21": "I-COLONY",
30
+ "22": "I-BLOCK",
31
+ "23": "I-CAMP",
32
+ "24": "I-POLE",
33
+ "25": "I-KHASRA",
34
+ "26": "I-FLOOR",
35
+ "27": "I-PLOT",
36
+ "28": "I-PINCODE",
37
+ "29": "I-CITY",
38
+ "30": "I-STATE"
39
+ },
40
+ "label2id": {
41
+ "O": 0,
42
+ "B-AREA": 1,
43
+ "B-SUBAREA": 2,
44
+ "B-HOUSE_NUMBER": 3,
45
+ "B-SECTOR": 4,
46
+ "B-GALI": 5,
47
+ "B-COLONY": 6,
48
+ "B-BLOCK": 7,
49
+ "B-CAMP": 8,
50
+ "B-POLE": 9,
51
+ "B-KHASRA": 10,
52
+ "B-FLOOR": 11,
53
+ "B-PLOT": 12,
54
+ "B-PINCODE": 13,
55
+ "B-CITY": 14,
56
+ "B-STATE": 15,
57
+ "I-AREA": 16,
58
+ "I-SUBAREA": 17,
59
+ "I-HOUSE_NUMBER": 18,
60
+ "I-SECTOR": 19,
61
+ "I-GALI": 20,
62
+ "I-COLONY": 21,
63
+ "I-BLOCK": 22,
64
+ "I-CAMP": 23,
65
+ "I-POLE": 24,
66
+ "I-KHASRA": 25,
67
+ "I-FLOOR": 26,
68
+ "I-PLOT": 27,
69
+ "I-PINCODE": 28,
70
+ "I-CITY": 29,
71
+ "I-STATE": 30
72
+ }
73
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b6d517ee453b37f6bfb28b2abbddf9bc5a9d2017e82842d5a903fd24781e32d6
3
+ size 1112346195
test_results.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "test_f1": 0.8,
3
+ "test_precision": 0.7931034482758621,
4
+ "test_recall": 0.8070175438596491,
5
+ "test_report": " precision recall f1-score support\n\n AREA 0.87 0.87 0.87 15\n CITY 1.00 1.00 1.00 24\n COLONY 0.00 0.00 0.00 2\n FLOOR 0.85 0.85 0.85 13\n GALI 0.75 0.67 0.71 9\nHOUSE_NUMBER 0.79 0.79 0.79 14\n KHASRA 0.75 0.82 0.78 11\n PINCODE 1.00 1.00 1.00 15\n PLOT 0.00 0.00 0.00 3\n SUBAREA 0.25 0.38 0.30 8\n\n micro avg 0.79 0.81 0.80 114\n macro avg 0.62 0.64 0.63 114\nweighted avg 0.80 0.81 0.80 114\n"
6
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "backend": "tokenizers",
3
+ "cls_token": "[CLS]",
4
+ "is_local": false,
5
+ "mask_token": "[MASK]",
6
+ "model_max_length": 1000000000000000019884624838656,
7
+ "pad_token": "[PAD]",
8
+ "sep_token": "[SEP]",
9
+ "tokenizer_class": "TokenizersBackend",
10
+ "unk_token": "[UNK]"
11
+ }
training_info.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_f1": 0.8868894601542417,
3
+ "best_precision": 0.8712121212121212,
4
+ "best_recall": 0.9031413612565445,
5
+ "epoch": 11,
6
+ "model_name": "ai4bharat/IndicBERTv2-SS",
7
+ "use_crf": true,
8
+ "learning_rate": 2e-05,
9
+ "crf_learning_rate": 0.001,
10
+ "lr_decay": 0.95,
11
+ "warmup_ratio": 0.1,
12
+ "train_samples": 1945,
13
+ "val_samples": 64
14
+ }