File size: 1,907 Bytes
2a7274f |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 |
---
license: cc-by-4.0
tags:
- generated_from_trainer
metrics:
- accuracy
- precision
- recall
- f1
model-index:
- name: hing-mbert-finetuned-TRAC-DS
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# hing-mbert-finetuned-TRAC-DS
This model is a fine-tuned version of [l3cube-pune/hing-mbert](https://huggingface.co/l3cube-pune/hing-mbert) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 1.9044
- Accuracy: 0.7010
- Precision: 0.6772
- Recall: 0.6723
- F1: 0.6740
## 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: 2.824279936868144e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 43
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:---------:|:------:|:------:|
| 0.837 | 1.0 | 1224 | 0.7640 | 0.6422 | 0.6377 | 0.6475 | 0.6277 |
| 0.6164 | 2.0 | 2448 | 0.8456 | 0.6724 | 0.6581 | 0.6623 | 0.6547 |
| 0.434 | 3.0 | 3672 | 1.0284 | 0.6969 | 0.6715 | 0.6771 | 0.6729 |
| 0.267 | 4.0 | 4896 | 1.5533 | 0.6912 | 0.6644 | 0.6675 | 0.6655 |
| 0.1542 | 5.0 | 6120 | 1.9044 | 0.7010 | 0.6772 | 0.6723 | 0.6740 |
### Framework versions
- Transformers 4.21.3
- Pytorch 1.12.1+cu113
- Datasets 2.4.0
- Tokenizers 0.12.1
|