lekhnathrijal commited on
Commit
a73d8a8
·
verified ·
1 Parent(s): c5471fb

ai-research-lab/bert-question-classifier

Browse files
Files changed (4) hide show
  1. README.md +33 -34
  2. config.json +18 -0
  3. model.safetensors +1 -1
  4. tokenizer.json +10 -1
README.md CHANGED
@@ -21,11 +21,11 @@ should probably proofread and complete it, then remove this comment. -->
21
 
22
  This model is a fine-tuned version of [google-bert/bert-base-uncased](https://huggingface.co/google-bert/bert-base-uncased) on the None dataset.
23
  It achieves the following results on the evaluation set:
24
- - Loss: 4.0243
25
- - Accuracy: 0.9374
26
- - Recall: 0.6628
27
- - Precision: 0.6533
28
- - F1: 0.6580
29
 
30
  ## Model description
31
 
@@ -44,9 +44,9 @@ More information needed
44
  ### Training hyperparameters
45
 
46
  The following hyperparameters were used during training:
47
- - learning_rate: 1e-06
48
- - train_batch_size: 8
49
- - eval_batch_size: 8
50
  - seed: 42
51
  - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
52
  - lr_scheduler_type: linear
@@ -57,32 +57,31 @@ The following hyperparameters were used during training:
57
 
58
  | Training Loss | Epoch | Step | Validation Loss | Accuracy | Recall | Precision | F1 |
59
  |:-------------:|:------:|:----:|:---------------:|:--------:|:------:|:---------:|:------:|
60
- | No log | 0.1916 | 100 | 5.6463 | 0.8581 | 0.2984 | 0.2578 | 0.2766 |
61
- | No log | 0.3831 | 200 | 5.4611 | 0.8740 | 0.3573 | 0.3246 | 0.3402 |
62
- | No log | 0.5747 | 300 | 5.1654 | 0.8954 | 0.3955 | 0.4201 | 0.4074 |
63
- | No log | 0.7663 | 400 | 4.9909 | 0.9097 | 0.4820 | 0.5037 | 0.4926 |
64
- | 5.3475 | 0.9579 | 500 | 4.8591 | 0.9167 | 0.5251 | 0.5433 | 0.5341 |
65
- | 5.3475 | 1.1494 | 600 | 4.7645 | 0.9201 | 0.5450 | 0.5622 | 0.5535 |
66
- | 5.3475 | 1.3410 | 700 | 4.6530 | 0.9242 | 0.5682 | 0.5855 | 0.5767 |
67
- | 5.3475 | 1.5326 | 800 | 4.5643 | 0.9266 | 0.5816 | 0.5992 | 0.5903 |
68
- | 5.3475 | 1.7241 | 900 | 4.4697 | 0.9286 | 0.5927 | 0.6107 | 0.6016 |
69
- | 4.7007 | 1.9157 | 1000 | 4.4025 | 0.9293 | 0.5965 | 0.6146 | 0.6054 |
70
- | 4.7007 | 2.1073 | 1100 | 4.3408 | 0.9302 | 0.6014 | 0.6197 | 0.6104 |
71
- | 4.7007 | 2.2989 | 1200 | 4.2894 | 0.9312 | 0.6067 | 0.6251 | 0.6158 |
72
- | 4.7007 | 2.4904 | 1300 | 4.2459 | 0.9307 | 0.6045 | 0.6227 | 0.6135 |
73
- | 4.7007 | 2.6820 | 1400 | 4.2056 | 0.9313 | 0.6092 | 0.6257 | 0.6173 |
74
- | 4.3569 | 2.8736 | 1500 | 4.1762 | 0.9320 | 0.6163 | 0.6286 | 0.6224 |
75
- | 4.3569 | 3.0651 | 1600 | 4.1439 | 0.9349 | 0.6408 | 0.6420 | 0.6414 |
76
- | 4.3569 | 3.2567 | 1700 | 4.1243 | 0.9352 | 0.6464 | 0.6426 | 0.6445 |
77
- | 4.3569 | 3.4483 | 1800 | 4.1034 | 0.9357 | 0.6501 | 0.6451 | 0.6476 |
78
- | 4.3569 | 3.6398 | 1900 | 4.0812 | 0.9364 | 0.6545 | 0.6490 | 0.6517 |
79
- | 4.192 | 3.8314 | 2000 | 4.0697 | 0.9363 | 0.6545 | 0.6480 | 0.6512 |
80
- | 4.192 | 4.0230 | 2100 | 4.0499 | 0.9370 | 0.6588 | 0.6521 | 0.6555 |
81
- | 4.192 | 4.2146 | 2200 | 4.0439 | 0.9370 | 0.6597 | 0.6517 | 0.6557 |
82
- | 4.192 | 4.4061 | 2300 | 4.0345 | 0.9371 | 0.6610 | 0.6521 | 0.6565 |
83
- | 4.192 | 4.5977 | 2400 | 4.0293 | 0.9373 | 0.6622 | 0.6529 | 0.6575 |
84
- | 4.1055 | 4.7893 | 2500 | 4.0262 | 0.9373 | 0.6622 | 0.6527 | 0.6574 |
85
- | 4.1055 | 4.9808 | 2600 | 4.0243 | 0.9374 | 0.6628 | 0.6533 | 0.6580 |
86
 
87
 
88
  ### Framework versions
 
21
 
22
  This model is a fine-tuned version of [google-bert/bert-base-uncased](https://huggingface.co/google-bert/bert-base-uncased) on the None dataset.
23
  It achieves the following results on the evaluation set:
24
+ - Loss: 1.5848
25
+ - Accuracy: 0.9719
26
+ - Recall: 0.8524
27
+ - Precision: 0.8409
28
+ - F1: 0.8466
29
 
30
  ## Model description
31
 
 
44
  ### Training hyperparameters
45
 
46
  The following hyperparameters were used during training:
47
+ - learning_rate: 1.5e-05
48
+ - train_batch_size: 4
49
+ - eval_batch_size: 4
50
  - seed: 42
51
  - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
52
  - lr_scheduler_type: linear
 
57
 
58
  | Training Loss | Epoch | Step | Validation Loss | Accuracy | Recall | Precision | F1 |
59
  |:-------------:|:------:|:----:|:---------------:|:--------:|:------:|:---------:|:------:|
60
+ | No log | 0.0959 | 100 | 5.2511 | 0.8846 | 0.4563 | 0.3862 | 0.4183 |
61
+ | No log | 0.1918 | 200 | 4.7859 | 0.9093 | 0.4991 | 0.5014 | 0.5002 |
62
+ | No log | 0.2876 | 300 | 4.3639 | 0.9228 | 0.6110 | 0.5702 | 0.5899 |
63
+ | No log | 0.3835 | 400 | 4.0136 | 0.9346 | 0.6752 | 0.6312 | 0.6525 |
64
+ | 4.7001 | 0.4794 | 500 | 3.7321 | 0.9383 | 0.6846 | 0.6535 | 0.6687 |
65
+ | 4.7001 | 0.5753 | 600 | 3.4591 | 0.9449 | 0.7146 | 0.6902 | 0.7022 |
66
+ | 4.7001 | 0.6711 | 700 | 3.3056 | 0.9425 | 0.6873 | 0.6825 | 0.6849 |
67
+ | 4.7001 | 0.7670 | 800 | 2.9697 | 0.9494 | 0.7404 | 0.7140 | 0.7270 |
68
+ | 4.7001 | 0.8629 | 900 | 2.7873 | 0.9511 | 0.7395 | 0.7271 | 0.7332 |
69
+ | 3.1425 | 0.9588 | 1000 | 2.6415 | 0.9551 | 0.7636 | 0.7481 | 0.7558 |
70
+ | 3.1425 | 1.0547 | 1100 | 2.4893 | 0.9581 | 0.7801 | 0.7640 | 0.7719 |
71
+ | 3.1425 | 1.1505 | 1200 | 2.3650 | 0.9600 | 0.7810 | 0.7798 | 0.7804 |
72
+ | 3.1425 | 1.2464 | 1300 | 2.1932 | 0.9645 | 0.8086 | 0.8026 | 0.8056 |
73
+ | 3.1425 | 1.3423 | 1400 | 2.1029 | 0.9651 | 0.8139 | 0.8042 | 0.8090 |
74
+ | 2.2748 | 1.4382 | 1500 | 2.0278 | 0.9666 | 0.8238 | 0.8120 | 0.8179 |
75
+ | 2.2748 | 1.5340 | 1600 | 1.9880 | 0.9660 | 0.8232 | 0.8067 | 0.8149 |
76
+ | 2.2748 | 1.6299 | 1700 | 1.8956 | 0.9679 | 0.8334 | 0.8167 | 0.8250 |
77
+ | 2.2748 | 1.7258 | 1800 | 1.8376 | 0.9686 | 0.8350 | 0.8225 | 0.8287 |
78
+ | 2.2748 | 1.8217 | 1900 | 1.7823 | 0.9690 | 0.8393 | 0.8232 | 0.8312 |
79
+ | 1.8388 | 1.9175 | 2000 | 1.7044 | 0.9707 | 0.8465 | 0.8338 | 0.8401 |
80
+ | 1.8388 | 2.0134 | 2100 | 1.6618 | 0.9709 | 0.8502 | 0.8331 | 0.8416 |
81
+ | 1.8388 | 2.1093 | 2200 | 1.6139 | 0.9716 | 0.8536 | 0.8372 | 0.8453 |
82
+ | 1.8388 | 2.2052 | 2300 | 1.5848 | 0.9719 | 0.8524 | 0.8409 | 0.8466 |
83
+ | 1.8388 | 2.3011 | 2400 | 1.5922 | 0.9716 | 0.8496 | 0.8397 | 0.8446 |
84
+ | 1.4663 | 2.3969 | 2500 | 1.5409 | 0.9719 | 0.8524 | 0.8409 | 0.8466 |
 
85
 
86
 
87
  ### Framework versions
config.json CHANGED
@@ -5,6 +5,24 @@
5
  ],
6
  "attention_probs_dropout_prob": 0.1,
7
  "classifier_dropout": null,
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
8
  "gradient_checkpointing": false,
9
  "hidden_act": "gelu",
10
  "hidden_dropout_prob": 0.1,
 
5
  ],
6
  "attention_probs_dropout_prob": 0.1,
7
  "classifier_dropout": null,
8
+ "custom_pipelines": {
9
+ "question-classifier": {
10
+ "default": {
11
+ "model": {
12
+ "pt": [
13
+ "ai-research-lab/bert-question-classifier",
14
+ "main"
15
+ ]
16
+ }
17
+ },
18
+ "impl": "classifier_pipeline.MultiTaskClassifierPipeline",
19
+ "pt": [
20
+ "AutoModelForSequenceClassification"
21
+ ],
22
+ "tf": [],
23
+ "type": "text"
24
+ }
25
+ },
26
  "gradient_checkpointing": false,
27
  "hidden_act": "gelu",
28
  "hidden_dropout_prob": 0.1,
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:03b42892a22e39f1b302840e5eb7aa350a6bc92d3e6a1522822c42bf65f0c740
3
  size 438057080
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1023f00f099b59e758a789fecd3eb91fa219b7c06f81aa9259cd9a72441ffc7a
3
  size 438057080
tokenizer.json CHANGED
@@ -6,7 +6,16 @@
6
  "strategy": "LongestFirst",
7
  "stride": 0
8
  },
9
- "padding": null,
 
 
 
 
 
 
 
 
 
10
  "added_tokens": [
11
  {
12
  "id": 0,
 
6
  "strategy": "LongestFirst",
7
  "stride": 0
8
  },
9
+ "padding": {
10
+ "strategy": {
11
+ "Fixed": 512
12
+ },
13
+ "direction": "Right",
14
+ "pad_to_multiple_of": null,
15
+ "pad_id": 0,
16
+ "pad_type_id": 0,
17
+ "pad_token": "[PAD]"
18
+ },
19
  "added_tokens": [
20
  {
21
  "id": 0,