Fix and improve REALM fine-tuning #15297
qqaatw
marked this pull request as ready for review 4 years ago
qqaatw
commented
on 2022-01-23
qqaatw
changed the title Prevent `block_emb` from sending to cuda in `RealmForOpenQA` model Fix and improve REALM 4 years ago
qqaatw
commented
on 2022-01-27
qqaatw
changed the title Fix and improve REALM Fix and improve REALM fine-tuning 4 years ago
qqaatw
changed the title Fix and improve REALM fine-tuning [WIP] Fix and improve REALM fine-tuning 4 years ago
qqaatw
force pushed
4 years ago
qqaatw
changed the title [WIP] Fix and improve REALM fine-tuning Fix and improve REALM fine-tuning 4 years ago
qqaatw
commented
on 2022-01-28
Draft
67bdd86c
Add test
00702686
Update src/transformers/models/realm/modeling_realm.py
ebd1d0d0
Apply suggestion
c2734aae
Add block_mask
3eef182f
Update
d6a38ced
Update
a3c99161
Add block_embedding_to
f8d6b3c9
Remove no_grad
acf76173
Use AutoTokenizer
bfca8df5
qqaatw
force pushed
to
bfca8df5
4 years ago
Remove model.to overridding
2c22cd34
Merge branch 'master' into fix_realm
823843b6
Assignees
No one assigned
Login to write a write a comment.
Login via GitHub