Downloads · 30 days
3
7% of all-time downloads
mdeberta-v3-base_mdeberta_EW-TT-PE_U0_S1_DROP1_E8_DEV93.0
pgajo/mdeberta-v3-base_mdeberta_EW-TT-PE_U0_S1_DROP1_E8_DEV93.0
mdeberta-v3-base_mdeberta_EW-TT-PE_U0_S1_DROP1_E8_DEV93.0 is a question answering model from pgajo. Use it when the input is a question plus a passage. It is set up for transformers.
All-time downloads
43
Public
Parameters
278M
1.1 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors1.1 GB · 98%
At a glance
- Task
- Question Answering
- Library
- transformers
- Model type
- deberta-v2
- Access
- Public
- Created
- Feb 11, 2024
- Updated
- Feb 11, 2024
- SHA
- b85579c0
Try a prompt
Model card
From the Hugging Face model README
Model description:
Model: microsoft/mdeberta-v3-base
Dataset: TASTEset
Unshuffled ratio: ['0']
Shuffled ratio: ['1']
Best exact match epoch: 8
Best exact match: 93.09
Best epoch: 8
Drop duplicates: ['1']
Max epochs = 10
Optimizer lr = 3e-05
Optimizer eps = 1e-08
Batch size = 8
Dataset path = pgajo/mdeberta_EW-TT-PE_U0_S1_DROP1
Results
| epoch | train_loss | train_f1 | train_exact | dev_loss | dev_f1 | dev_exact | test_loss | test_f1 | test_exact |
|---|---|---|---|---|---|---|---|---|---|
| 1 | 1.95 | 48.6 | 40.98 | 0.46 | 85.03 | 81.22 | 0 | 0 | 0 |
| 2 | 0.36 | 88.68 | 85.63 | 0.39 | 90.41 | 89.23 | 0 | 0 | 0 |
| 3 | 0.25 | 91.68 | 89.5 | 0.36 | 90.08 | 88.12 | 0 | 0 | 0 |
| 4 | 0.13 | 95.32 | 94.4 | 0.29 | 91.88 | 90.61 | 0 | 0 | 0 |
| 5 | 0.09 | 97 | 96.27 | 0.3 | 93.72 | 92.54 | 0 | 0 | 0 |
| 6 | 0.09 | 97.04 | 96.34 | 0.37 | 91.44 | 89.78 | 0 | 0 | 0 |
| 7 | 0.07 | 97.2 | 96.61 | 0.29 | 92.68 | 91.99 | 0 | 0 | 0 |
| 8 | 0.05 | 98.34 | 98.06 | 0.33 | 93.5 | 93.09 | 0 | 0 | 0 |
| 9 | 0.03 | 98.67 | 98.55 | 0.35 | 93.67 | 91.99 | 0 | 0 | 0 |
| 10 | 0.02 | 99.33 | 98.96 | 0.4 | 93.54 | 92.54 | 0 | 0 | 0 |
- Task
- Question Answering
- Library
- transformers
- Type
- deberta-v2
- Created
- Feb 11, 2024
- Updated
- Feb 11, 2024