Update README.md
Browse files
README.md
CHANGED
|
@@ -1,12 +1,16 @@
|
|
| 1 |
---
|
| 2 |
library_name: transformers
|
| 3 |
-
|
|
|
|
| 4 |
---
|
| 5 |
|
| 6 |
See the run here: https://wandb.ai/stadeltom-com/huggingface/runs/0kmc811n?nw=nwuserstadeltom
|
| 7 |
|
| 8 |
Best f1 was 90.1%, accuracy 90.1%, precision 87.2%, recall 95% (at the best f1)
|
| 9 |
|
|
|
|
|
|
|
|
|
|
| 10 |
# Model Card for Model ID
|
| 11 |
|
| 12 |
<!-- Provide a quick summary of what the model is/does. -->
|
|
|
|
| 1 |
---
|
| 2 |
library_name: transformers
|
| 3 |
+
datasets:
|
| 4 |
+
- tom-010/squad_v2_with_answerable
|
| 5 |
---
|
| 6 |
|
| 7 |
See the run here: https://wandb.ai/stadeltom-com/huggingface/runs/0kmc811n?nw=nwuserstadeltom
|
| 8 |
|
| 9 |
Best f1 was 90.1%, accuracy 90.1%, precision 87.2%, recall 95% (at the best f1)
|
| 10 |
|
| 11 |
+
The idea was to freeze the body for a few thausend samples, then train on squadv2 answerability.
|
| 12 |
+
But this had no effect, model is same quality as without the freezing.
|
| 13 |
+
|
| 14 |
# Model Card for Model ID
|
| 15 |
|
| 16 |
<!-- Provide a quick summary of what the model is/does. -->
|