admin/lerobot
1
0
Fork 0
mirror of https://github.com/huggingface/lerobot.git synced 2026-07-19 16:02:11 +00:00
Code Issues Packages Projects Releases Wiki Activity

feat(record): Integrate RobotProcessor into recording loop and update policy handling

Browse Source 
- Added support for RobotProcessor in the record_loop function to enhance data processing capabilities.
- Updated the logic to reset both policy and processor when provided, ensuring proper state management.
- Modified action prediction to utilize the processor, improving the overall functionality of the recording process.
- Adjusted the save_checkpoint function to include preprocessor state saving, enhancing checkpointing capabilities.
This commit is contained in:
AdilZouitine
2025-07-17 13:10:15 +02:00
committed by Steven Palma
parent 670a278cbc
commit 4b24f94225
6 changed files with 36 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/lerobot/scripts/train.py
+1 -1
View File
@@ -244,7 +244,7 @@ def train(cfg: TrainPipelineConfig):
if cfg.save_checkpoint and is_saving_step:
logging.info(f"Checkpoint policy after step {step}")
checkpoint_dir = get_step_checkpoint_dir(cfg.output_dir, cfg.steps, step)
save_checkpoint(checkpoint_dir, step, cfg, policy, optimizer, lr_scheduler)
save_checkpoint(checkpoint_dir, step, cfg, policy, optimizer, lr_scheduler, preprocessor)
update_last_checkpoint(checkpoint_dir)
if wandb_logger:
wandb_logger.log_policy(checkpoint_dir)