File size: 2,158 Bytes
8200dfd |
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 |
---
language:
- zh
- ko
base_model: facebook/mbart-large-50-many-to-many-mmt
tags:
- generated_from_trainer
metrics:
- bleu
model-index:
- name: zhko_mbartLarge_19p_run3
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. -->
# zhko_mbartLarge_19p_run3
This model is a fine-tuned version of [facebook/mbart-large-50-many-to-many-mmt](https://huggingface.co/facebook/mbart-large-50-many-to-many-mmt) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 1.4430
- Bleu: 17.067
- Gen Len: 14.7398
## 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: 3.5e-05
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- total_eval_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 1500
- num_epochs: 10
### Training results
| Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len |
|:-------------:|:-----:|:-----:|:---------------:|:-------:|:-------:|
| 1.957 | 0.45 | 1250 | 1.7573 | 12.1283 | 15.4356 |
| 1.6568 | 0.9 | 2500 | 1.5793 | 14.3528 | 14.9632 |
| 1.4964 | 1.35 | 3750 | 1.4806 | 16.0215 | 14.9365 |
| 1.2204 | 1.79 | 5000 | 1.4682 | 16.5619 | 14.8621 |
| 1.1678 | 2.24 | 6250 | 1.4430 | 16.9926 | 14.7721 |
| 0.9448 | 2.69 | 7500 | 1.4615 | 17.0327 | 14.5641 |
| 0.9083 | 3.14 | 8750 | 1.4744 | 17.863 | 14.6946 |
| 0.7389 | 3.59 | 10000 | 1.5129 | 17.3256 | 14.7824 |
### Framework versions
- Transformers 4.34.1
- Pytorch 2.1.0+cu121
- Datasets 2.14.6
- Tokenizers 0.14.1
|