File size: 1,900 Bytes
e0fa7af 68ffb01 e0fa7af 68ffb01 e0fa7af 68ffb01 e0fa7af |
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 |
---
license: cc-by-nc-4.0
base_model: facebook/nllb-200-distilled-600M
tags:
- generated_from_trainer
metrics:
- bleu
model-index:
- name: nllb-fr-dje
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. -->
# nllb-fr-dje
This model is a fine-tuned version of [facebook/nllb-200-distilled-600M](https://huggingface.co/facebook/nllb-200-distilled-600M) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.2285
- Bleu: 29.6795
## 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: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 8
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 8
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Bleu |
|:-------------:|:-----:|:----:|:---------------:|:-------:|
| 2.8423 | 1.0 | 508 | 1.8073 | 20.0301 |
| 1.7444 | 2.0 | 1017 | 1.4882 | 25.3031 |
| 1.4902 | 3.0 | 1526 | 1.3685 | 27.3325 |
| 1.3532 | 4.0 | 2035 | 1.3004 | 28.0924 |
| 1.2749 | 5.0 | 2544 | 1.2632 | 29.0171 |
| 1.2153 | 6.0 | 3053 | 1.2432 | 29.3495 |
| 1.1793 | 7.0 | 3562 | 1.2323 | 29.5367 |
| 1.1548 | 7.99 | 4064 | 1.2285 | 29.6795 |
### Framework versions
- Transformers 4.35.0
- Pytorch 2.0.0
- Datasets 2.1.0
- Tokenizers 0.14.1
|