Hugging Face's logo

exentai
/
indicner-tamil-ExentAI

Token Classification
Transformers
TensorBoard
Safetensors
bert
Generated from Trainer
Model card Files
xet
 Metrics Community
exentai commited on
Commit
bd1bbb7
·
verified ·
1 Parent(s): a9073dd

Fine-tuned on Srilankan-Tamil-NER v2 (10K, 10 epochs)

Browse files
Files changed (9) hide show
  1. README.md +87 -0
  2. config.json +55 -0
  3. loss_curves.png +0 -0
  4. model.safetensors +3 -0
  5. runs/Jul16_04-49-40_22dbd0354e6a/events.out.tfevents.1784177380.22dbd0354e6a.965.0 +3 -0
  6. runs/Jul16_04-49-40_22dbd0354e6a/events.out.tfevents.1784178627.22dbd0354e6a.965.1 +3 -0
  7. tokenizer.json +0 -0
  8. tokenizer_config.json +20 -0
  9. training_args.bin +3 -0
README.md ADDED
@@ -0,0 +1,87 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: transformers
3
+ license: mit
4
+ base_model: ai4bharat/IndicNER
5
+ tags:
6
+ - generated_from_trainer
7
+ metrics:
8
+ - precision
9
+ - recall
10
+ - f1
11
+ - accuracy
12
+ model-index:
13
+ - name: indicner-tamil-ExentAI
14
+ results: []
15
+ ---
16
+
17
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
18
+ should probably proofread and complete it, then remove this comment. -->
19
+
20
+ # indicner-tamil-ExentAI
21
+
22
+ This model is a fine-tuned version of [ai4bharat/IndicNER](https://huggingface.co/ai4bharat/IndicNER) on an unknown dataset.
23
+ It achieves the following results on the evaluation set:
24
+ - Loss: 0.1632
25
+ - Precision: 0.6005
26
+ - Recall: 0.7003
27
+ - F1: 0.6466
28
+ - Accuracy: 0.9624
29
+ - F1 Per: 0.6897
30
+ - Precision Per: 0.6481
31
+ - Recall Per: 0.7368
32
+ - F1 Loc: 0.7113
33
+ - Precision Loc: 0.6646
34
+ - Recall Loc: 0.7652
35
+ - F1 Org: 0.4625
36
+ - Precision Org: 0.4190
37
+ - Recall Org: 0.5161
38
+
39
+ ## Model description
40
+
41
+ More information needed
42
+
43
+ ## Intended uses & limitations
44
+
45
+ More information needed
46
+
47
+ ## Training and evaluation data
48
+
49
+ More information needed
50
+
51
+ ## Training procedure
52
+
53
+ ### Training hyperparameters
54
+
55
+ The following hyperparameters were used during training:
56
+ - learning_rate: 6.857179151838835e-05
57
+ - train_batch_size: 64
58
+ - eval_batch_size: 64
59
+ - seed: 42
60
+ - optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
61
+ - lr_scheduler_type: cosine
62
+ - lr_scheduler_warmup_steps: 0.13646586401400382
63
+ - num_epochs: 10
64
+ - mixed_precision_training: Native AMP
65
+
66
+ ### Training results
67
+
68
+ | Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy | F1 Per | Precision Per | Recall Per | F1 Loc | Precision Loc | Recall Loc | F1 Org | Precision Org | Recall Org |
69
+ |:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|:------:|:-------------:|:----------:|:------:|:-------------:|:----------:|:------:|:-------------:|:----------:|
70
+ | 0.4201 | 1.0 | 132 | 0.2198 | 0.2407 | 0.3604 | 0.2886 | 0.9252 | 0.2605 | 0.2394 | 0.2857 | 0.3803 | 0.3033 | 0.5097 | 0.0286 | 0.0244 | 0.0346 |
71
+ | 0.1555 | 2.0 | 264 | 0.1483 | 0.4512 | 0.5869 | 0.5102 | 0.9498 | 0.4578 | 0.4136 | 0.5126 | 0.6416 | 0.5649 | 0.7425 | 0.2059 | 0.1791 | 0.2421 |
72
+ | 0.1010 | 3.0 | 396 | 0.1445 | 0.4723 | 0.6263 | 0.5385 | 0.9531 | 0.4713 | 0.4109 | 0.5525 | 0.6770 | 0.6155 | 0.7521 | 0.2826 | 0.2294 | 0.3679 |
73
+ | 0.0600 | 4.0 | 528 | 0.1481 | 0.5467 | 0.6408 | 0.5900 | 0.9585 | 0.5315 | 0.4937 | 0.5756 | 0.7049 | 0.6745 | 0.7382 | 0.3760 | 0.3214 | 0.4528 |
74
+ | 0.0411 | 5.0 | 660 | 0.1546 | 0.5680 | 0.6605 | 0.6108 | 0.9611 | 0.5448 | 0.4860 | 0.6197 | 0.7189 | 0.6922 | 0.7479 | 0.4163 | 0.3766 | 0.4654 |
75
+ | 0.0298 | 6.0 | 792 | 0.1730 | 0.5534 | 0.6929 | 0.6154 | 0.9591 | 0.5714 | 0.5116 | 0.6471 | 0.7100 | 0.6442 | 0.7908 | 0.4120 | 0.3639 | 0.4748 |
76
+ | 0.0199 | 7.0 | 924 | 0.1812 | 0.5778 | 0.6842 | 0.6265 | 0.9612 | 0.5787 | 0.5277 | 0.6408 | 0.7305 | 0.6885 | 0.7779 | 0.4131 | 0.3656 | 0.4748 |
77
+ | 0.0170 | 8.0 | 1056 | 0.1914 | 0.5863 | 0.7005 | 0.6383 | 0.9608 | 0.5806 | 0.5265 | 0.6471 | 0.7398 | 0.6943 | 0.7918 | 0.4454 | 0.3937 | 0.5126 |
78
+ | 0.0138 | 9.0 | 1188 | 0.1975 | 0.5824 | 0.6999 | 0.6358 | 0.9608 | 0.5833 | 0.5281 | 0.6513 | 0.7334 | 0.6870 | 0.7865 | 0.4472 | 0.3929 | 0.5189 |
79
+ | 0.0120 | 10.0 | 1320 | 0.1998 | 0.5800 | 0.6976 | 0.6334 | 0.9607 | 0.5805 | 0.5236 | 0.6513 | 0.7316 | 0.6871 | 0.7822 | 0.4453 | 0.3901 | 0.5189 |
80
+
81
+
82
+ ### Framework versions
83
+
84
+ - Transformers 5.12.1
85
+ - Pytorch 2.11.0+cu128
86
+ - Datasets 4.0.0
87
+ - Tokenizers 0.22.2
config.json ADDED
@@ -0,0 +1,55 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_cross_attention": false,
3
+ "architectures": [
4
+ "BertForTokenClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.1,
7
+ "bos_token_id": null,
8
+ "classifier_dropout": null,
9
+ "directionality": "bidi",
10
+ "dtype": "float32",
11
+ "eos_token_id": null,
12
+ "finetuning_task": "ner",
13
+ "gradient_checkpointing": false,
14
+ "hidden_act": "gelu",
15
+ "hidden_dropout_prob": 0.1,
16
+ "hidden_size": 768,
17
+ "id2label": {
18
+ "0": "O",
19
+ "1": "B-PER",
20
+ "2": "I-PER",
21
+ "3": "B-LOC",
22
+ "4": "I-LOC",
23
+ "5": "B-ORG",
24
+ "6": "I-ORG"
25
+ },
26
+ "initializer_range": 0.02,
27
+ "intermediate_size": 3072,
28
+ "is_decoder": false,
29
+ "label2id": {
30
+ "B-LOC": 3,
31
+ "B-ORG": 5,
32
+ "B-PER": 1,
33
+ "I-LOC": 4,
34
+ "I-ORG": 6,
35
+ "I-PER": 2,
36
+ "O": 0
37
+ },
38
+ "layer_norm_eps": 1e-12,
39
+ "max_position_embeddings": 512,
40
+ "model_type": "bert",
41
+ "num_attention_heads": 12,
42
+ "num_hidden_layers": 12,
43
+ "pad_token_id": 0,
44
+ "pooler_fc_size": 768,
45
+ "pooler_num_attention_heads": 12,
46
+ "pooler_num_fc_layers": 3,
47
+ "pooler_size_per_head": 128,
48
+ "pooler_type": "first_token_transform",
49
+ "position_embedding_type": "absolute",
50
+ "tie_word_embeddings": true,
51
+ "transformers_version": "5.12.1",
52
+ "type_vocab_size": 2,
53
+ "use_cache": false,
54
+ "vocab_size": 105879
55
+ }
loss_curves.png ADDED
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:546638a1ff895a9c81123a8ce905b233e73868b2ea3d145897dfe654923ffec7
3
+ size 667108188
runs/Jul16_04-49-40_22dbd0354e6a/events.out.tfevents.1784177380.22dbd0354e6a.965.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a538556a65d9a0afe7a9bfb25b84df3c78d12de114415b59ba4177388871167d
3
+ size 20354
runs/Jul16_04-49-40_22dbd0354e6a/events.out.tfevents.1784178627.22dbd0354e6a.965.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0e497d77433188575e192936e9eb78a37df76ef66268b9550c62aa185eac7023
3
+ size 1043
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,20 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "backend": "tokenizers",
3
+ "cls_token": "[CLS]",
4
+ "do_lower_case": true,
5
+ "is_local": false,
6
+ "keep_accents": true,
7
+ "local_files_only": false,
8
+ "mask_token": "[MASK]",
9
+ "max_length": 512,
10
+ "model_max_length": 512,
11
+ "pad_token": "[PAD]",
12
+ "sep_token": "[SEP]",
13
+ "stride": 0,
14
+ "strip_accents": null,
15
+ "tokenize_chinese_chars": true,
16
+ "tokenizer_class": "BertTokenizer",
17
+ "truncation_side": "right",
18
+ "truncation_strategy": "longest_first",
19
+ "unk_token": "[UNK]"
20
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:84d7da9c6d8fe4e12e80c7398e28a8679a0936315359354f270f17d4e2d62562
3
+ size 5201