tangger/issacdataengine
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

rich annotations & update open-pi fsdp explanations

Browse Source
This commit is contained in:
Leon998
2026-03-18 13:59:52 +08:00
parent 814f3c3526
commit 4934c4794e
11 changed files with 349 additions and 32 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
policy/openpi-InternData-A1/docs/training.md
View File

@@ -40,7 +40,7 @@ python scripts/download_paligemma.py
You may adjust other training parameters based on your available GPUs and training budget:
- `num_train_steps`: Total number of training steps
- `num_workers`: Number of data loading workers
- `fsdp_devices`: Number of GPUs per node
- `fsdp_devices`: Number of GPUs used for FSDP per node to distribute model parameters, gradients, and optimizer states across devices for reduced memory usage
- `batch_size`: Batch size per GPU
- `save_interval`: Checkpoint saving interval (in steps)