File size: 2,122 Bytes
49a0e9c |
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 |
---
tags:
- generated_from_trainer
metrics:
- bleu
model-index:
- name: t5-small-ENG2BASH-NL2BASH-customv1-customv2
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. -->
# t5-small-ENG2BASH-NL2BASH-customv1-customv2
This model is a fine-tuned version of [alexsha/t5-small-ENG2BASH-NL2BASH-customv1](https://huggingface.co/alexsha/t5-small-ENG2BASH-NL2BASH-customv1) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3117
- Bleu: 62.6753
- Gen Len: 17.36
## 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: 0.0002
- train_batch_size: 5
- eval_batch_size: 5
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|
| No log | 1.0 | 40 | 1.3531 | 21.0375 | 12.04 |
| No log | 2.0 | 80 | 0.6775 | 56.9739 | 16.88 |
| No log | 3.0 | 120 | 0.5087 | 61.3993 | 17.2 |
| No log | 4.0 | 160 | 0.4243 | 58.4332 | 17.2 |
| No log | 5.0 | 200 | 0.3642 | 61.2412 | 17.44 |
| No log | 6.0 | 240 | 0.3367 | 63.1132 | 17.56 |
| No log | 7.0 | 280 | 0.3181 | 63.1132 | 17.56 |
| No log | 8.0 | 320 | 0.3140 | 62.9769 | 17.44 |
| No log | 9.0 | 360 | 0.3136 | 63.861 | 17.32 |
| No log | 10.0 | 400 | 0.3117 | 62.6753 | 17.36 |
### Framework versions
- Transformers 4.27.0.dev0
- Pytorch 1.13.1
- Datasets 2.9.0
- Tokenizers 0.13.2
|