Downloads · 30 days
6
13% of all-time downloads
mbert_EW-TT-PE_U0_S1_Tingredient_P0.75_DROP1_mbert_E9_DEV89.0
pgajo/mbert_EW-TT-PE_U0_S1_Tingredient_P0.75_DROP1_mbert_E9_DEV89.0
mbert_EW-TT-PE_U0_S1_Tingredient_P0.75_DROP1_mbert_E9_DEV89.0 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
45
Public
Parameters
177M
709 MB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors709 MB · 99%
At a glance
- Task
- Question Answering
- Library
- transformers
- Model type
- bert
- Access
- Public
- Created
- Feb 13, 2024
- Updated
- Feb 13, 2024
- SHA
- 9052ed78
Try a prompt
Model card
From the Hugging Face model README
Model description:
Model: bert-base-multilingual-cased
Dataset: TASTEset
Unshuffled ratio: ['0']
Shuffled ratio: ['1']
Best exact match epoch: 9
Best exact match: 89.29
Best epoch: 9
Drop duplicates: ['1']
Max epochs = 10
Optimizer lr = 3e-05
Optimizer eps = 1e-08
Batch size = 32
Dataset path = pgajo/EW-TT-PE_U0_S1_Tingredient_P0.75_DROP1_mbert
Results
| epoch | train_loss | train_f1 | train_exact | dev_loss | dev_f1 | dev_exact | test_loss | test_f1 | test_exact |
|---|---|---|---|---|---|---|---|---|---|
| 1 | 3.23 | 10.64 | 2.55 | 2.56 | 18.12 | 8.52 | 0 | 0 | 0 |
| 2 | 1.2 | 59 | 48.62 | 0.59 | 83.96 | 75.27 | 0 | 0 | 0 |
| 3 | 0.37 | 88.86 | 83.61 | 0.46 | 91.16 | 85.16 | 0 | 0 | 0 |
| 4 | 0.17 | 94.22 | 91.18 | 0.48 | 90.52 | 85.44 | 0 | 0 | 0 |
| 5 | 0.09 | 97.37 | 95.8 | 0.5 | 89.31 | 83.79 | 0 | 0 | 0 |
| 6 | 0.06 | 98.07 | 96.76 | 0.47 | 91.89 | 89.01 | 0 | 0 | 0 |
| 7 | 0.04 | 98.67 | 97.59 | 0.53 | 92.25 | 87.36 | 0 | 0 | 0 |
| 8 | 0.04 | 98.93 | 97.93 | 0.48 | 92.88 | 89.01 | 0 | 0 | 0 |
| 9 | 0.03 | 99.31 | 98.9 | 0.51 | 93.68 | 89.29 | 0 | 0 | 0 |
| 10 | 0.01 | 99.57 | 99.24 | 0.5 | 94.13 | 89.29 | 0 | 0 | 0 |
- Task
- Question Answering
- Library
- transformers
- Type
- bert
- Created
- Feb 13, 2024
- Updated
- Feb 13, 2024