File size: 2,002 Bytes
67b05dc |
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 |
---
library_name: transformers
tags:
- generated_from_trainer
model-index:
- name: Beit_PhoBart
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. -->
# Beit_PhoBart
This model is a fine-tuned version of [](https://huggingface.co/) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 1.5865
- Bleu-4: 0.0954
## 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: 5e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- gradient_accumulation_steps: 2
- 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: linear
- lr_scheduler_warmup_steps: 300
- num_epochs: 100
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Bleu-4 |
|:-------------:|:-----:|:----:|:---------------:|:------:|
| No log | 1.0 | 236 | 1.7574 | 0.0672 |
| 1.9578 | 2.0 | 472 | 1.6226 | 0.0748 |
| 1.5495 | 3.0 | 708 | 1.5523 | 0.0796 |
| 1.3763 | 4.0 | 944 | 1.4998 | 0.0821 |
| 1.3763 | 5.0 | 1180 | 1.4815 | 0.0881 |
| 1.2355 | 6.0 | 1416 | 1.4835 | 0.0875 |
| 1.1027 | 7.0 | 1652 | 1.4790 | 0.0926 |
| 0.9865 | 8.0 | 1888 | 1.5128 | 0.0925 |
| 0.884 | 9.0 | 2124 | 1.5453 | 0.0956 |
| 0.884 | 10.0 | 2360 | 1.5865 | 0.0954 |
### Framework versions
- Transformers 4.46.2
- Pytorch 2.5.1+cu121
- Datasets 3.1.0
- Tokenizers 0.20.3
|