bert-30M-uncased-classification-fqa

This model is a fine-tuned version of vietgpt/bert-30M-uncased on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 0.8275
  • Accuracy: 0.9538

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

  • learning_rate: 2e-05
  • train_batch_size: 16
  • eval_batch_size: 16
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 50

Training results

Training Loss Epoch Step Validation Loss Accuracy
No log 1.0 110 5.2109 0.0051
No log 2.0 220 5.2109 0.0
No log 3.0 330 5.1842 0.0
No log 4.0 440 5.0942 0.0256
5.141 5.0 550 4.8190 0.0615
5.141 6.0 660 4.5678 0.1128
5.141 7.0 770 4.3158 0.2205
5.141 8.0 880 4.0876 0.3333
5.141 9.0 990 3.8774 0.4410
4.3509 10.0 1100 3.6849 0.5179
4.3509 11.0 1210 3.5020 0.6103
4.3509 12.0 1320 3.3304 0.6718
4.3509 13.0 1430 3.1628 0.7026
3.3612 14.0 1540 3.0022 0.7692
3.3612 15.0 1650 2.8528 0.7897
3.3612 16.0 1760 2.7042 0.8051
3.3612 17.0 1870 2.5607 0.8205
3.3612 18.0 1980 2.4300 0.8359
2.5735 19.0 2090 2.3086 0.8462
2.5735 20.0 2200 2.1888 0.8462
2.5735 21.0 2310 2.0762 0.8667
2.5735 22.0 2420 1.9736 0.8718
1.9651 23.0 2530 1.8741 0.8718
1.9651 24.0 2640 1.7797 0.9026
1.9651 25.0 2750 1.6888 0.9077
1.9651 26.0 2860 1.6092 0.9128
1.9651 27.0 2970 1.5296 0.9128
1.5059 28.0 3080 1.4617 0.9179
1.5059 29.0 3190 1.3951 0.9179
1.5059 30.0 3300 1.3362 0.9179
1.5059 31.0 3410 1.2752 0.9179
1.1862 32.0 3520 1.2268 0.9179
1.1862 33.0 3630 1.1743 0.9179
1.1862 34.0 3740 1.1333 0.9231
1.1862 35.0 3850 1.0918 0.9282
1.1862 36.0 3960 1.0535 0.9385
0.9586 37.0 4070 1.0206 0.9385
0.9586 38.0 4180 0.9899 0.9333
0.9586 39.0 4290 0.9638 0.9385
0.9586 40.0 4400 0.9418 0.9436
0.8104 41.0 4510 0.9163 0.9436
0.8104 42.0 4620 0.8983 0.9436
0.8104 43.0 4730 0.8811 0.9487
0.8104 44.0 4840 0.8657 0.9538
0.8104 45.0 4950 0.8551 0.9538
0.7164 46.0 5060 0.8448 0.9538
0.7164 47.0 5170 0.8384 0.9487
0.7164 48.0 5280 0.8326 0.9538
0.7164 49.0 5390 0.8289 0.9538
0.6757 50.0 5500 0.8275 0.9538

Framework versions

  • Transformers 4.37.1
  • Pytorch 2.1.0+cu121
  • Datasets 2.16.1
  • Tokenizers 0.15.1
Downloads last month
19
Safetensors
Model size
34.6M params
Tensor type
F32
·
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Model tree for thangvip/bert-30M-uncased-classification-fqa

Finetuned
(6)
this model