Downloads · 30 days
10
3% of all-time downloads
deberta-large_spell_5k_2_p3
stuartmesham/deberta-large_spell_5k_2_p3
deberta-large_spell_5k_2_p3 is a token classification model from stuartmesham. Use it when you need labels on individual words, such as names. It is set up for transformers.
should probably proofread and complete it, then remove this comment. --
All-time downloads
316
Public
Repo size
3.3 GB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.bin1.6 GB · 100%
At a glance
- Task
- Token Classification
- Library
- transformers
- Model type
- deberta
- Access
- Public
- Created
- Oct 24, 2022
- Updated
- Oct 24, 2022
- SHA
- 4c8cb156
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. -->
deberta-large_spell_5k_2_p3
This model is a fine-tuned version of model_saves/deberta-large_spell_5k_2_p2 on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.4141
- Accuracy: 0.9416
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: 1e-05
- train_batch_size: 128
- eval_batch_size: 128
- seed: 52
- distributed_type: multi-GPU
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: constant
- num_epochs: 15.0
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| No log | 1.0 | 268 | 0.4141 | 0.9416 |
| No log | 2.0 | 536 | 0.4367 | 0.9412 |
| No log | 3.0 | 804 | 0.4807 | 0.9400 |
| 0.255 | 4.0 | 1072 | 0.5355 | 0.9398 |
Framework versions
- Transformers 4.21.0.dev0
- Pytorch 1.11.0+cu113
- Datasets 2.3.2
- Tokenizers 0.12.1
- Task
- Token Classification
- Library
- transformers
- Type
- deberta
- Created
- Oct 24, 2022
- Updated
- Oct 24, 2022