Downloads · 30 days
15
25% of all-time downloads
REPEAT622_5wangchanberta-base-att-spm-uncased
Tonjk/REPEAT622_5wangchanberta-base-att-spm-uncased
REPEAT622_5wangchanberta-base-att-spm-uncased is a fill-mask model from Tonjk. Use it when you need the model to fill a missing word. It is set up for transformers.
should probably proofread and complete it, then remove this comment. --
All-time downloads
59
Public
Repo size
850 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.bin425 MB · 100%
At a glance
- Task
- Fill-Mask
- Library
- transformers
- Model type
- camembert
- Access
- Public
- Created
- Dec 31, 2022
- Updated
- Dec 31, 2022
- SHA
- c04d9445
Try a prompt
Model card
From the Hugging Face model README
<!-- 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. -->
REPEAT622_5wangchanberta-base-att-spm-uncased
This model is a fine-tuned version of airesearch/wangchanberta-base-att-spm-uncased on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.0898
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: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 0.4527 | 1.0 | 9531 | 0.3250 |
| 0.2652 | 2.0 | 19062 | 0.6947 |
| 0.1967 | 3.0 | 28593 | 0.8256 |
| 0.1667 | 4.0 | 38124 | 0.7543 |
| 0.1515 | 5.0 | 47655 | 1.0898 |
Framework versions
- Transformers 4.15.0
- Pytorch 1.13.0+cu116
- Datasets 1.17.0
- Tokenizers 0.10.3
- Task
- Fill-Mask
- Library
- transformers
- Type
- camembert
- Created
- Dec 31, 2022
- Updated
- Dec 31, 2022