Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

bump transformers and update attention class map name #1023

Merged
merged 16 commits into from
Jan 3, 2024
Merged

bump transformers and update attention class map name #1023

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
16 commits
Select commit Hold shift + click to select a range
984ec2d
bump transformers and update attention class map name
winglian Dec 30, 2023
387cdc9
also run the tests in docker
winglian Dec 31, 2023
a91badf
add mixtral e2e smoke test
winglian Dec 31, 2023
2d46145
fix base name for docker image in test
winglian Dec 31, 2023
b9a9b14
mixtral lora doesn't seem to work, at least check qlora
winglian Dec 31, 2023
ff1277e
add testcase for mixtral w sample packing
winglian Jan 1, 2024
c5f108b
check monkeypatch for flash attn multipack
winglian Jan 1, 2024
b9fcdbf
also run the e2e tests in docker
winglian Jan 1, 2024
99cbfb1
use all gpus to run tests in docker ci
winglian Jan 2, 2024
6882e7c
use privileged mode too for docker w gpus
winglian Jan 2, 2024
b4d47e0
rename the docker e2e actions for gh ci
winglian Jan 2, 2024
44fa9ff
set privileged mode for docker and update mixtral model self attn check
winglian Jan 2, 2024
658a660
use fp16/bf16 for mixtral w fa2
winglian Jan 2, 2024
2e23f38
skip e2e tests on docker w gpus for now
winglian Jan 2, 2024
7a3f24b
tests to validate mistral and mixtral patches
winglian Jan 3, 2024
e29b20f
fix rel import
winglian Jan 3, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion requirements.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,7 +2,7 @@
auto-gptq==0.5.1
packaging
peft==0.6.0
transformers==4.36.2
transformers @ git+https://github.com/huggingface/transformers.git@3cefac1d974db5e2825a0cb2b842883a628be7a0
tokenizers==0.15.0
bitsandbytes>=0.41.1
accelerate==0.24.1
Expand Down
2 changes: 1 addition & 1 deletion src/axolotl/monkeypatch/mixtral/__init__.py
View file
Open in desktop
winglian marked this conversation as resolved.
Show resolved Hide resolved
Original file line number Diff line number Diff line change
Expand Up @@ -17,6 +17,6 @@ def replace_mixtral_attn_with_multipack_flash_attn():
transformers.models.mixtral.modeling_mixtral.MixtralModel.forward = (
mixtral_model_forward
)
transformers.models.mixtral.modeling_mixtral.MISTRAL_ATTENTION_CLASSES[
transformers.models.mixtral.modeling_mixtral.MIXTRAL_ATTENTION_CLASSES[
"flash_attention_2"
] = MixtralMultipackFlashAttention2