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

Migrate OH CLIP (roberta-clip) training to torch.compile #1507

Open
wants to merge 1 commit into
base: main
Choose a base branch
from
Open

Migrate OH CLIP (roberta-clip) training to torch.compile #1507

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
24 changes: 12 additions & 12 deletions examples/contrastive-image-text/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -96,7 +96,7 @@ Finally, we can run the example script to train the model.
Run the following command for single-device training:

```bash
python run_clip.py \
PT_HPU_LAZY_MODE=0 python run_clip.py \
--output_dir ./clip-roberta-finetuned \
--model_name_or_path ./clip-roberta \
--data_dir $PWD/data \
Expand All @@ -112,14 +112,13 @@ python run_clip.py \
--overwrite_output_dir \
--save_strategy epoch \
--use_habana \
--use_lazy_mode \
--use_hpu_graphs_for_training \
--use_hpu_graphs_for_inference \
--gaudi_config_name Habana/clip \
--throughput_warmup_steps 3 \
--dataloader_num_workers 16 \
--bf16 \
--trust_remote_code
--trust_remote_code \
--torch_compile_backend=hpu_backend \
--torch_compile
```


Expand All @@ -128,6 +127,7 @@ python run_clip.py \
Run the following command for distributed training:

```bash
PT_HPU_LAZY_MODE=0 \
python ../gaudi_spawn.py --world_size 8 --use_mpi run_clip.py \
--output_dir ./clip-roberta-finetuned \
--model_name_or_path ./clip-roberta \
Expand All @@ -144,16 +144,15 @@ python ../gaudi_spawn.py --world_size 8 --use_mpi run_clip.py \
--overwrite_output_dir \
--save_strategy epoch \
--use_habana \
--use_lazy_mode \
--use_hpu_graphs_for_inference \
--gaudi_config_name Habana/clip \
--throughput_warmup_steps 3 \
--dataloader_num_workers 16 \
--mediapipe_dataloader \
--use_hpu_graphs_for_training \
--bf16 \
--distribution_strategy fast_ddp \
--trust_remote_code
--trust_remote_code \
--torch_compile_backend=hpu_backend \
--torch_compile
```

> `--mediapipe_dataloader` only works on Gaudi2.
Expand All @@ -164,6 +163,7 @@ python ../gaudi_spawn.py --world_size 8 --use_mpi run_clip.py \
Run the following command for training with DeepSpeed:

```bash
PT_HPU_LAZY_MODE=0 \
python ../gaudi_spawn.py --world_size 8 --use_deepspeed run_clip.py \
--output_dir ./clip-roberta-finetuned \
--model_name_or_path ./clip-roberta \
Expand All @@ -180,12 +180,12 @@ python ../gaudi_spawn.py --world_size 8 --use_deepspeed run_clip.py \
--overwrite_output_dir \
--save_strategy epoch \
--use_habana \
--use_lazy_mode \
--use_hpu_graphs_for_inference \
--gaudi_config_name Habana/clip \
--throughput_warmup_steps 3 \
--deepspeed path_to_my_deepspeed_config \
--trust_remote_code
--trust_remote_code \
--torch_compile_backend=hpu_backend \
--torch_compile
```

You can look at the [documentation](https://huggingface.co/docs/optimum/habana/usage_guides/deepspeed) for more information about how to use DeepSpeed in Optimum Habana.
Expand Down