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

[Example] Refactor and Polish Cifar10-DeepSpeed Code Example. #843

Merged
merged 3 commits into from
Jan 26, 2024
Merged

[Example] Refactor and Polish Cifar10-DeepSpeed Code Example. #843

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
4b6e087
Polish and Refactor Cifar10 Code Example
keli-wen Jan 5, 2024
bc67dc9
Merge branch 'master' into PolishCifar10Example
tjruwase Jan 23, 2024
e53dfd5
fix typos
conglongli Jan 26, 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
23 changes: 12 additions & 11 deletions training/cifar/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,21 +1,22 @@
Thanks Gopi Kumar for contributing this example, demonstrating how to apply DeepSpeed to CIFAR-10 model.

cifar10_tutorial.py
`cifar10_tutorial.py`
Baseline CIFAR-10 model.

cifar10_deepspeed.py
`cifar10_deepspeed.py`
DeepSpeed applied CIFAR-10 model.

ds_config.json
DeepSpeed configuration file.

run_ds.sh
`run_ds.sh`
Script for running DeepSpeed applied model.

run_ds_moe.sh
`run_ds_moe.sh`
Script for running DeepSpeed model with Mixture of Experts (MoE) integration.

* To run baseline CIFAR-10 model - "python cifar10_tutorial.py"
* To run DeepSpeed CIFAR-10 model - "bash run_ds.sh"
* To run DeepSpeed CIFAR-10 model with Mixture of Experts (MoE) - "bash run_ds_moe.sh"
* To run with different data type (default='fp16') and zero stages (default=0) - "bash run_ds.sh --dtype={fp16|bf16} --stage={0|1|2|3}"
`run_ds_prmode.sh`
keli-wen marked this conversation as resolved.
Show resolved Hide resolved
Script for running DeepSpeed model with Pyramid Residual MoE (PR-MoE) integration.

* To run baseline CIFAR-10 model - `python cifar10_tutorial.py`
* To run DeepSpeed CIFAR-10 model - `bash run_ds.sh`
* To run DeepSpeed CIFAR-10 model with Mixture of Experts (MoE) - `bash run_ds_moe.sh`
* To run DeepSpeed CIFAR-10 model with Pyramid Residual MoE (PR-MoE) - `bash run_ds_prmode.sh`
keli-wen marked this conversation as resolved.
Show resolved Hide resolved
* To run with different data type (default=`fp16`) and zero stages (default=`0`) - `bash run_ds.sh --dtype={fp16|bf16} --stage={0|1|2|3}`
Loading
Loading