LeRobot-SO101-MultiTaskDiT-task2-all_bs128_s30000

LeRobot SO101 MultiTaskDiT task2-all_bs128_s30000

Summary

This repository contains the final checkpoint for a MultiTask DiT policy trained on aswinkumar99/task2-all for SO101 sponge pick-and-place experiments.

Dataset meaning: Task 2: Multiple Sponges - No Distractors (all layouts).

This model was trained with the LeRobot multi_task_dit policy and diffusion objective. It is not a fine-tune from a published base checkpoint.

Training Setup

• Dataset repo: aswinkumar99/task2-all
• Local dataset root during training: /home/riftuser/datasets_combined/aswinkumar99/task2-all
• Output directory during training: /home/riftuser/outputs_matrix/multi_task_dit/task2-all_bs128_s30000
• Batch size: 128
• Training steps: 30000
• Checkpoint save frequency: 5000
• Data loader workers: 8
• WandB project: so101-layout-generalization
• GPU: NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition
• Python: CPython 3.12.13
• CUDA: 12.9
• Training start: 2026-04-24T09:48:49.302378+00:00
• Training end: 2026-04-24T18:02:09
• Approximate training duration: 8h 13m 19s
• Objective: diffusion
• Noise scheduler: DDPM
• Horizon: 32
• Action steps predicted: 24
• Observation steps: 2
• Vision encoder: openai/clip-vit-base-patch16
• Text encoder: openai/clip-vit-base-patch16
• Hidden dim: 512
• Number of transformer layers: 4

Exact Training Command

lerobot-train \
  --dataset.repo_id=aswinkumar99/task2-all \
  --dataset.root=/home/riftuser/datasets_combined/aswinkumar99/task2-all \
  --dataset.video_backend=torchcodec \
  --output_dir=/home/riftuser/outputs_matrix/multi_task_dit/task2-all_bs128_s30000 \
  --job_name=multi_task_dit_task2-all_bs128 \
  --batch_size=128 \
  --steps=30000 \
  --log_freq=200 \
  --save_freq=5000 \
  --save_checkpoint=true \
  --num_workers=8 \
  --wandb.enable=true \
  --wandb.project=so101-layout-generalization \
  --wandb.mode=online \
  --wandb.disable_artifact=true \
  --policy.type=multi_task_dit \
  --policy.device=cuda \
  --policy.push_to_hub=false \
  --policy.use_amp=true \
  --policy.horizon=32 \
  --policy.n_action_steps=24 \
  --policy.n_obs_steps=2 \
  --policy.num_layers=4 \
  --policy.hidden_dim=512 \
  --policy.num_heads=8 \
  --policy.dropout=0.1 \
  --policy.timestep_embed_dim=256 \
  --policy.use_rope=true \
  --policy.use_positional_encoding=false \
  --policy.objective=diffusion \
  --policy.noise_scheduler_type=DDPM \
  --policy.num_train_timesteps=100 \
  --policy.optimizer_lr=2e-5 \
  --policy.vision_encoder_lr_multiplier=0.1 \
  --policy.vision_encoder_name=openai/clip-vit-base-patch16 \
  --policy.text_encoder_name=openai/clip-vit-base-patch16 \
  --policy.image_crop_shape=[224,224] \
  --policy.image_crop_is_random=true

Repository Contents

• pretrained_model/: final downloadable model artifacts for inference/loading
• training_state/: optimizer, RNG, scheduler/state, and step information for resuming or auditability

Creator

Aswinkumar

• Website: aswinkumar.me
• Hugging Face repo: https://huggingface.co/aswinkumar99/LeRobot-SO101-MultiTaskDiT-task2-all_bs128_s30000