Downloads · 30 days
3
7% of all-time downloads
mdeberta_EW-TT-PE_U0_S1_Tingredient_DROP1_mdeberta_E4_DEV96.0
pgajo/mdeberta_EW-TT-PE_U0_S1_Tingredient_DROP1_mdeberta_E4_DEV96.0
mdeberta_EW-TT-PE_U0_S1_Tingredient_DROP1_mdeberta_E4_DEV96.0 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
41
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 12, 2024
- Updated
- Feb 12, 2024
- SHA
- 7b053a9e
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: 4
Best exact match: 96.43
Best epoch: 4
Drop duplicates: ['1']
Max epochs = 10
Optimizer lr = 3e-05
Optimizer eps = 1e-08
Batch size = 8
Dataset path = pgajo/EW-TT-PE_U0_S1_Tingredient_DROP1_mdeberta
Results
| epoch | train_loss | train_f1 | train_exact | dev_loss | dev_f1 | dev_exact | test_loss | test_f1 | test_exact |
|---|---|---|---|---|---|---|---|---|---|
| 1 | 2.23 | 40.25 | 31.82 | 0.36 | 91.75 | 88.46 | 0 | 0 | 0 |
| 2 | 0.27 | 92.71 | 90.01 | 0.19 | 95.5 | 93.68 | 0 | 0 | 0 |
| 3 | 0.1 | 96.91 | 95.73 | 0.15 | 97.05 | 95.88 | 0 | 0 | 0 |
| 4 | 0.06 | 98.38 | 97.45 | 0.12 | 96.95 | 96.43 | 0 | 0 | 0 |
| 5 | 0.05 | 98.66 | 97.66 | 0.15 | 96.43 | 93.68 | 0 | 0 | 0 |
| 6 | 0.06 | 98.48 | 97.66 | 0.14 | 95.86 | 94.51 | 0 | 0 | 0 |
| 7 | 0.04 | 98.85 | 98.28 | 0.17 | 95.83 | 95.05 | 0 | 0 | 0 |
- Task
- Question Answering
- Library
- transformers
- Type
- deberta-v2
- Created
- Feb 12, 2024
- Updated
- Feb 12, 2024