 Dmitry Baranchuk
|
6095f58681
Deep distributed prompt tuning (#42)
|
3 years ago |
|
Dmitry Baranchuk
|
11a424837f
integrate mixed-8bit model (#39)
|
3 years ago |
|
Dmitry Baranchuk
|
04a2b6f5e3
Support various backend dtypes & async serialization (#38)
|
3 years ago |
 Artem Chumachenko
|
d989b94614
Pack of Inference Changes (#37)
|
3 years ago |
|
Dmitry Baranchuk
|
6573076883
Sequential and parallel forward / backward (#36)
|
3 years ago |
|
Artem Chumachenko
|
6ee942e915
Add GenerationMixin class (#29)
|
3 years ago |
 justheuristic
|
e2711a033b
Add automated tests (#23)
|
3 years ago |
|
Dmitry Baranchuk
|
f5463812ad
Shallow prompt tuning (#22)
|
3 years ago |
|
dbaranchuk
|
21e1f42f04
mv set_requires_grad to remote_model
|
3 years ago |
|
dbaranchuk
|
79280c4371
refactoring
|
3 years ago |
|
dbaranchuk
|
6bffeff0a1
fix
|
3 years ago |
|
dbaranchuk
|
b3cc9e0d99
add LM head for DistributedBloomCausalLM
|
3 years ago |
|
dbaranchuk
|
df42822f26
LM head for CausalLM & chunked forward
|
3 years ago |
|
justheuristic
|
1c68670d06
it works
|
3 years ago |
|
justheuristic
|
88c1bf9896
black-isort
|
3 years ago |
|
justheuristic
|
4695071ad2
WIP: make DistributedBloom compliant with HF interface
|
3 years ago |
|
justheuristic
|
e32208c954
black-isort
|
3 years ago |
|
justheuristic
|
4ad845bce3
black-isort
|
3 years ago |
|
Dmitry Baranchuk
|
e66ab6f1f2
design interface & refactoring
|
3 years ago |
 Dmitry Baranchuk
|
d969172208
set requires_grad=False, lm_layer -> h @ word_embeddings, rm lm_layer from comverted_model
|
3 years ago |
|
justheuristic
|
331591c915
less intrusive warnings
|
3 years ago |
|
justheuristic
|
9c492bbe8c
Infer prefix by defaukt
|
3 years ago |
|
justheuristic
|
19ae71e8fc
from_pretrained
|
3 years ago |
|
justheuristic
|
471e47c0f5
black-isort
|
3 years ago |
|
justheuristic
|
7d68f6b9a4
fix model creation
|
3 years ago |
|
justheuristic
|
5849cea28c
prototype remote sequential
|
3 years ago |
