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

[Doc][Example] Fine-tune vicuna-13b-v1.3 with LightningTrainer + DeepSpeed #37016

Merged
merged 19 commits into from
Jul 25, 2023
Merged

[Doc][Example] Fine-tune vicuna-13b-v1.3 with LightningTrainer + DeepSpeed #37016

merged 19 commits into from
Jul 25, 2023

Conversation

woshiyyya
Copy link
Member

@woshiyyya woshiyyya commented Jul 1, 2023 •
edited
Loading

Why are these changes needed?

This PR addes a new end-to-end LLM fine-tuning example for LightningTrainer DeepSpeed Integration. We fine-tuned a vicuna-13b-v1.3 model on a code generation dataset with LightningTrainer on 16 A10G GPUs. The major contents includes:

  • Data pre-processing and prompting
  • Fine-tuning with DeepSpeed ZeRO-3
  • Distributed Checkpointing
  • Single-GPU Model Initialization and Inference
  • Case study using Aviary

Rendered Doc: https://anyscale-ray--37016.com.readthedocs.build/en/37016/ray-air/examples/vicuna_13b_lightning_deepspeed_finetune.html

Finished release test: https://buildkite.com/ray-project/release-tests-pr/builds/46242#01898512-1c77-4bcf-b7c4-c2437ae7e1bc

Related issue number

Resources:

Screenshot 2023-07-03 at 6 47 40 PM

Checks

  • I've signed off every commit(by using the -s flag, i.e., git commit -s) in this PR.
  • I've run scripts/format.sh to lint the changes in this PR.
  • I've included any doc changes needed for https://docs.ray.io/en/master/.
    • I've added any new APIs to the API Reference. For example, if I added a
      method in Tune, I've added it in doc/source/tune/api/ under the
      corresponding .rst file.
  • I've made sure the tests are passing. Note that there might be a few flaky tests, see the recent failures at https://flakey-tests.ray.io/
  • Testing Strategy
    • Unit tests
    • Release tests
    • This PR is not tested :(

woshiyyya added 5 commits June 30, 2023 19:25
@woshiyyya
init
22c81ac 
Signed-off-by: woshiyyya <xiaoyunxuan1998@gmail.com>
@woshiyyya
add release tests
f6f8185 
Signed-off-by: woshiyyya <xiaoyunxuan1998@gmail.com>
@woshiyyya
add text in ipynb
54ec91f 
Signed-off-by: woshiyyya <xiaoyunxuan1998@gmail.com>
@woshiyyya
add references
b2c6f9d 
Signed-off-by: woshiyyya <xiaoyunxuan1998@gmail.com>
@woshiyyya
rename and add to toc.yml
7813867 
Signed-off-by: woshiyyya <xiaoyunxuan1998@gmail.com>
@woshiyyya woshiyyya marked this pull request as ready for review July 4, 2023 04:40
woshiyyya added 2 commits July 5, 2023 12:01
@woshiyyya
polished wordings
1236d10 
Signed-off-by: woshiyyya <xiaoyunxuan1998@gmail.com>
@woshiyyya
polish outputs
00a8525 
Signed-off-by: woshiyyya <xiaoyunxuan1998@gmail.com>
woshiyyya added 5 commits July 5, 2023 13:00
@woshiyyya
Merge remote-tracking branch 'upstream/master' into doc/vicuna_deepsp…
c34450b 
…eed_lightning_example
@woshiyyya
hide one cell
0ae9dda 
Signed-off-by: woshiyyya <xiaoyunxuan1998@gmail.com>
@woshiyyya
exclude from doc test
b01e36d 
Signed-off-by: woshiyyya <xiaoyunxuan1998@gmail.com>
@woshiyyya
Merge remote-tracking branch 'upstream/master' into doc/vicuna_deepsp…
139b083 
…eed_lightning_example
@woshiyyya
remove preprocessor
75e323b 
Signed-off-by: woshiyyya <xiaoyunxuan1998@gmail.com>
matthewdeng
matthewdeng approved these changes Jul 20, 2023
View reviewed changes
Copy link
Contributor

@matthewdeng matthewdeng left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is really great! Mostly minor grammar nits.

doc/source/_toc.yml Outdated Show resolved Hide resolved
doc/source/ray-air/examples/vicuna_13b_lightning_deepspeed_finetune.ipynb Outdated Show resolved Hide resolved
doc/source/ray-air/examples/vicuna_13b_lightning_deepspeed_finetune.ipynb Outdated Show resolved Hide resolved
doc/source/ray-air/examples/vicuna_13b_lightning_deepspeed_finetune.ipynb Outdated Show resolved Hide resolved
doc/source/ray-air/examples/vicuna_13b_lightning_deepspeed_finetune.ipynb Outdated Show resolved Hide resolved
doc/source/ray-air/examples/vicuna_13b_lightning_deepspeed_finetune.ipynb Outdated Show resolved Hide resolved
...ples/vicuna_13b_lightning_deepspeed_finetuning/vicuna_13b_lightning_deepspeed_finetune.ipynb Outdated
@@ -0,0 +1 @@
../../../doc/source/ray-air/examples/vicuna_13b_lightning_deepspeed_finetune.ipynb
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Do you know why this is red?

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is a sym link file. I think GH renders this in a different color.

release/air_examples/vicuna_13b_lightning_deepspeed_finetuning/vicuna_13b_deepspeed_env.yaml Outdated Show resolved Hide resolved
release/air_examples/vicuna_13b_lightning_deepspeed_finetuning/vicuna_13b_deepspeed_env.yaml Outdated Show resolved Hide resolved
doc/source/ray-air/examples/vicuna_13b_lightning_deepspeed_finetune.ipynb Outdated Show resolved Hide resolved
woshiyyya and others added 4 commits July 20, 2023 10:51
@woshiyyya @matthewdeng
Apply suggestions from code review
8f76166 
Co-authored-by: matthewdeng <matthew.j.deng@gmail.com>
Signed-off-by: Yunxuan Xiao <xiaoyunxuan1998@gmail.com>
@woshiyyya
address comments
54ab9aa 
Signed-off-by: woshiyyya <xiaoyunxuan1998@gmail.com>
@woshiyyya
Merge remote-tracking branch 'upstream/master' into doc/vicuna_deepsp…
ad44a7f 
…eed_lightning_example
@woshiyyya
pin typing-extenstions
1ee92f0 
Signed-off-by: woshiyyya <xiaoyunxuan1998@gmail.com>
@woshiyyya
clear SIGUSR1 message
255fe4f 
Signed-off-by: woshiyyya <xiaoyunxuan1998@gmail.com>
matthewdeng
matthewdeng approved these changes Jul 24, 2023
View reviewed changes
doc/source/_toc.yml
@@ -141,6 +141,8 @@ parts:
title: "Torch Data Prefetching Benchmark"
- file: train/examples/pytorch/pytorch_resnet_finetune
title: "PyTorch Finetuning ResNet Example"
- file: train/examples/lightning/vicuna_13b_lightning_deepspeed_finetune
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can you add this to the Example Gallery as well?

@woshiyyya
add it into example gallery
242b75a 
Signed-off-by: woshiyyya <xiaoyunxuan1998@gmail.com>
@matthewdeng
Update doc/source/ray-overview/examples.rst
265f8a6 
Signed-off-by: matthewdeng <matthew.j.deng@gmail.com>
@matthewdeng matthewdeng merged commit 6c059c4 into ray-project:master Jul 25, 2023
NripeshN pushed a commit to NripeshN/ray that referenced this pull request Aug 15, 2023
@woshiyyya @matthewdeng @NripeshN
[Doc][Example] Fine-tune vicuna-13b-v1.3 with LightningTrainer + De…
0b4e573 
…epSpeed (ray-project#37016)

Signed-off-by: woshiyyya <xiaoyunxuan1998@gmail.com>
Signed-off-by: Yunxuan Xiao <xiaoyunxuan1998@gmail.com>
Signed-off-by: matthewdeng <matthew.j.deng@gmail.com>
Co-authored-by: matthewdeng <matthew.j.deng@gmail.com>
Signed-off-by: NripeshN <nn2012@hw.ac.uk>
arvind-chandra pushed a commit to lmco/ray that referenced this pull request Aug 31, 2023
@woshiyyya @matthewdeng @arvind-chandra
[Doc][Example] Fine-tune vicuna-13b-v1.3 with LightningTrainer + De…
7a9f3e0 
…epSpeed (ray-project#37016)

Signed-off-by: woshiyyya <xiaoyunxuan1998@gmail.com>
Signed-off-by: Yunxuan Xiao <xiaoyunxuan1998@gmail.com>
Signed-off-by: matthewdeng <matthew.j.deng@gmail.com>
Co-authored-by: matthewdeng <matthew.j.deng@gmail.com>
Signed-off-by: e428265 <arvind.chandramouli@lmco.com>
vymao pushed a commit to vymao/ray that referenced this pull request Oct 11, 2023
@woshiyyya @matthewdeng
[Doc][Example] Fine-tune vicuna-13b-v1.3 with LightningTrainer + De…
fe0d046 
…epSpeed (ray-project#37016)

Signed-off-by: woshiyyya <xiaoyunxuan1998@gmail.com>
Signed-off-by: Yunxuan Xiao <xiaoyunxuan1998@gmail.com>
Signed-off-by: matthewdeng <matthew.j.deng@gmail.com>
Co-authored-by: matthewdeng <matthew.j.deng@gmail.com>
Signed-off-by: Victor <vctr.y.m@example.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@matthewdeng matthewdeng matthewdeng approved these changes

@richardliaw richardliaw Awaiting requested review from richardliaw

@gjoliver gjoliver Awaiting requested review from gjoliver

@krfricke krfricke Awaiting requested review from krfricke

@xwjiang2010 xwjiang2010 Awaiting requested review from xwjiang2010

@amogkam amogkam Awaiting requested review from amogkam

@Yard1 Yard1 Awaiting requested review from Yard1

@maxpumperla maxpumperla Awaiting requested review from maxpumperla

Assignees

@matthewdeng matthewdeng

@richardliaw richardliaw

@Yard1 Yard1

Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
4 participants
@woshiyyya @matthewdeng @richardliaw @Yard1