metadata
license: apache-2.0
base_model: t5-small
tags:
- generated_from_trainer
datasets:
- flores
metrics:
- bleu
model-index:
- name: meta-translation-chinese-english-model
results:
- task:
name: Sequence-to-sequence Language Modeling
type: text2text-generation
dataset:
name: flores
type: flores
config: zho_Hans-eng_Latn
split: dev
args: zho_Hans-eng_Latn
metrics:
- name: Bleu
type: bleu
value: 0.4467
meta-translation-chinese-english-model
This model is a fine-tuned version of t5-small on the flores dataset. It achieves the following results on the evaluation set:
- Loss: 3.8438
- Bleu: 0.4467
- Gen Len: 18.815
- Exact Match: 0.0
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: 16
- eval_batch_size: 16
- 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 | Exact Match |
|---|---|---|---|---|---|---|
| No log | 1.0 | 50 | 3.9743 | 0.1872 | 15.755 | 0.0 |
| No log | 2.0 | 100 | 3.9226 | 0.196 | 17.525 | 0.0 |
| No log | 3.0 | 150 | 3.8931 | 0.3008 | 18.075 | 0.0 |
| No log | 4.0 | 200 | 3.8743 | 0.4408 | 18.83 | 0.0 |
| No log | 5.0 | 250 | 3.8628 | 0.3987 | 18.83 | 0.0 |
| No log | 6.0 | 300 | 3.8546 | 0.4146 | 18.905 | 0.0 |
| No log | 7.0 | 350 | 3.8498 | 0.4069 | 18.87 | 0.0 |
| No log | 8.0 | 400 | 3.8464 | 0.4022 | 18.85 | 0.0 |
| No log | 9.0 | 450 | 3.8445 | 0.4432 | 18.815 | 0.0 |
| 4.0794 | 10.0 | 500 | 3.8438 | 0.4467 | 18.815 | 0.0 |
Framework versions
- Transformers 4.42.4
- Pytorch 2.3.1+cu121
- Datasets 2.21.0
- Tokenizers 0.19.1