Downloads · 30 days
8
1% of all-time downloads
microsoft-deberta-v3-large_ner_wikiann
Gladiator/microsoft-deberta-v3-large_ner_wikiann
microsoft-deberta-v3-large_ner_wikiann is a token classification model from Gladiator. Use it when you need labels on individual words, such as names. It is set up for transformers. The card lists the license as mit.
should probably proofread and complete it, then remove this comment. --
All-time downloads
551
Public
Repo size
10.4 GB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.bin1.7 GB · 99%
At a glance
- Task
- Token Classification
- Library
- transformers
- License
- mit
- Model type
- deberta-v2
- Access
- Public
- Created
- Dec 10, 2022
- Updated
- Dec 10, 2022
- SHA
- 96b7f32c
Try a prompt
Datasets
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. -->
microsoft-deberta-v3-large_ner_wikiann
This model is a fine-tuned version of microsoft/deberta-v3-large on the wikiann dataset. It achieves the following results on the evaluation set:
- Loss: 0.3108
- Precision: 0.8557
- Recall: 0.8738
- F1: 0.8647
- Accuracy: 0.9406
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: cosine
- num_epochs: 5
Training results
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|---|---|---|---|---|---|---|---|
| 0.3005 | 1.0 | 1250 | 0.2462 | 0.8205 | 0.8400 | 0.8301 | 0.9294 |
| 0.1931 | 2.0 | 2500 | 0.2247 | 0.8448 | 0.8630 | 0.8538 | 0.9386 |
| 0.1203 | 3.0 | 3750 | 0.2341 | 0.8468 | 0.8693 | 0.8579 | 0.9403 |
| 0.0635 | 4.0 | 5000 | 0.2948 | 0.8596 | 0.8745 | 0.8670 | 0.9411 |
| 0.0451 | 5.0 | 6250 | 0.3108 | 0.8557 | 0.8738 | 0.8647 | 0.9406 |
Framework versions
- Transformers 4.24.0
- Pytorch 1.13.0+cu117
- Datasets 2.7.1
- Tokenizers 0.13.2
- Task
- Token Classification
- Library
- transformers
- Type
- deberta-v2
- License
- mit
- Created
- Dec 10, 2022
- Updated
- Dec 10, 2022