File size: 3,208 Bytes
f3daa6e | 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 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 | ---
library_name: transformers
license: mit
base_model: microsoft/mdeberta-v3-base
tags:
- generated_from_trainer
model-index:
- name: bengali_qa_microsoft_model
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. -->
# bengali_qa_microsoft_model
This model is a fine-tuned version of [microsoft/mdeberta-v3-base](https://huggingface.co/microsoft/mdeberta-v3-base) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 2.7354
- Exact Match: 47.8571
- F1 Score: 64.4066
## 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: 8
- eval_batch_size: 8
- seed: 3407
- gradient_accumulation_steps: 8
- total_train_batch_size: 64
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- training_steps: 50
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Exact Match | F1 Score |
|:-------------:|:------:|:----:|:---------------:|:-----------:|:--------:|
| 6.6257 | 0.0053 | 1 | 6.6659 | 0.1504 | 23.7019 |
| 6.6148 | 0.0107 | 2 | 6.5377 | 0.7519 | 36.0663 |
| 6.5217 | 0.0160 | 3 | 6.3226 | 1.9549 | 43.4507 |
| 6.1229 | 0.0214 | 4 | 5.6034 | 3.7594 | 56.8030 |
| 5.7015 | 0.0267 | 5 | 5.3006 | 5.7143 | 56.8976 |
| 5.3474 | 0.0321 | 6 | 5.1196 | 10.9023 | 56.6826 |
| 5.2375 | 0.0374 | 7 | 4.8802 | 16.9925 | 57.8356 |
| 5.067 | 0.0428 | 8 | 4.6180 | 19.8496 | 57.7216 |
| 4.7947 | 0.0481 | 9 | 4.3354 | 22.7820 | 58.3592 |
| 4.4271 | 0.0534 | 10 | 4.0533 | 26.0902 | 58.7345 |
| 4.3096 | 0.0588 | 11 | 3.7947 | 31.5789 | 60.0493 |
| 4.1219 | 0.0641 | 12 | 3.5726 | 35.2632 | 61.0031 |
| 3.9806 | 0.0695 | 13 | 3.4024 | 38.4962 | 62.6457 |
| 3.653 | 0.0748 | 14 | 3.2782 | 41.7293 | 63.9058 |
| 3.474 | 0.0802 | 15 | 3.1569 | 43.9850 | 65.0367 |
| 3.3639 | 0.0855 | 16 | 3.0200 | 45.4887 | 64.9433 |
| 3.2411 | 0.0908 | 17 | 2.8749 | 46.4662 | 64.8581 |
| 3.0711 | 0.0962 | 18 | 2.7349 | 47.8195 | 65.1471 |
| 3.0726 | 0.1015 | 19 | 2.6103 | 48.4962 | 64.6205 |
| 2.9879 | 0.1069 | 20 | 2.4996 | 49.2481 | 64.7963 |
| 2.9237 | 0.1122 | 21 | 2.3950 | 50.6767 | 65.0445 |
### Framework versions
- Transformers 4.46.3
- Pytorch 2.4.0
- Datasets 3.1.0
- Tokenizers 0.20.3
|