dodox/dppo
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Fix WandB configuration for proper logging

Browse Source 
- Configure DPPO_WANDB_ENTITY environment variable in dev script
- Update README with clear WandB setup instructions
- Remove wandb=null to enable logging when credentials are set
This commit is contained in:
ys1087@partner.kit.edu 2025-08-27 12:23:43 +02:00
parent e8e7233d98
commit d43a9e2b3c
2 changed files with 18 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
README.md
View File

@ -119,14 +119,22 @@ tail -f logs/dppo_<job_id>.out
### Configuration
Before running experiments, set your WandB credentials:
**Required for WandB logging**: Set your credentials as environment variables before submitting jobs:
```bash
# Set these in your shell before running experiments
export WANDB_API_KEY=<your_api_key>
export WANDB_ENTITY=<your_username_or_team>
export DPPO_WANDB_ENTITY=<your_username_or_team>
# Then submit jobs normally
./submit_job.sh dev
```
Or disable WandB by adding `wandb=null` to your python command.
**Alternative**: Modify the SLURM scripts directly to set your credentials:
- Edit `slurm/run_dppo_dev.sh` and replace `"your_wandb_username"` with your actual WandB username
- Uncomment and set `WANDB_API_KEY` in the scripts
**Disable WandB**: Add `wandb=null` to python commands if you don't want logging.
### Repository Changes

13
slurm/run_dppo_dev.sh
View File

@ -14,11 +14,12 @@
# Load required modules
module load devel/cuda/12.4
# Set environment variables
# Set environment variables for WandB
export WANDB_MODE=online
export WANDB_PROJECT=dppo_dev_test
# export WANDB_API_KEY=<your_api_key> # TODO: Set your API key
# export WANDB_ENTITY=<your_entity> # TODO: Set your entity (username or team)
export DPPO_WANDB_ENTITY=${DPPO_WANDB_ENTITY:-"your_wandb_username"} # TODO: Set your WandB username
# TODO: Set your WandB API key as environment variable:
# export WANDB_API_KEY=<your_api_key>
# Default paths (can be overridden by environment)
export DPPO_DATA_DIR=${DPPO_DATA_DIR:-$SLURM_SUBMIT_DIR/data}
@ -41,11 +42,11 @@ echo "PyTorch version: $(python -c 'import torch; print(torch.__version__)')"
echo "CUDA available: $(python -c 'import torch; print(torch.cuda.is_available())')"
echo ""
# Run a quick pre-training test with reduced epochs (disable WandB for dev test)
# Run a quick pre-training test with reduced epochs
# Note: Will only log to WandB if WANDB_API_KEY and DPPO_WANDB_ENTITY are properly set
python script/run.py --config-name=pre_diffusion_mlp \
--config-dir=cfg/gym/pretrain/hopper-medium-v2 \
train.n_epochs=2 \
train.save_model_freq=1 \
wandb=null
train.save_model_freq=1
echo "Dev test completed!"