Skip to content
Robotics

rimur01/diffusion_so101_assignment4_50k

diffusion policy for robotic manipulator

2 min read

The model implements the Diffusion Policy algorithm (an approach where robot control is framed as generating a sequence of smooth actions) for an so_follower type robotic arm. Trained on a dataset of 50 episodes, the model controls object gripping and placement into a target zone based on frames from two cameras and joint states. It was built using the LeRobot library for imitation learning in real-world robotics. The solution is suitable for researchers and developers testing open-source robotic arms.

Key features

  • Diffusion Policy: generates multi-step smooth trajectories for contact manipulation tasks
  • Multimodal input: processes robot state (6 degrees of freedom) and images from two cameras (cam0, cam1) in 3x480x640 format
  • Robot compatibility: native integration for the so_follower robot type
  • LeRobot tool: training and running via the Hugging Face LeRobot ecosystem version 0.6.0

What you can do

  • Run the policy on a real so_follower robotic arm using the LeRobot command-line utility
  • Test the task of picking up and moving objects to a target area
  • Configure inference parameters capturing video streams from two USB cameras
  • Use the dataset rimur01/so101-assignment1-50-new to fine-tune your own policy variations

Example

Running the ready-made inference policy on the robot via the command line:

lerobot-rollout \
  --strategy.type=base \
  --robot.type=so_follower \
  --robot.port= \
  --robot.cameras="{ : {type: opencv, index_or_path: , width: 640, height: 480, fps: 30}, : {type: opencv, index_or_path: , width: 640, height: 480, fps: 30}}" \
  --policy.path=rimur01/diffusion_so101_assignment4_50k \
  --task="Pick up an object and place it in a target area" \
  --duration=60
Read moreDetails
License: apache-2.0
Training steps: 50,000 with batch size 1
Optimizer: adam (learning rate: 0.0001, seed: 1000)
Dataset: 50 episodes, 15,000 frames at 15 FPS
Discuss Telegram 0 22

Will it run?

Runs in F32

Fits without quantisation.

  • F32weights 0.1 GB · ready files0.6 GB
  • BF16weights 0.1 GB0.6 GB
  • FP8 / INT8recommendedweights 0.0 GB0.5 GB
  • 4-bitweights 0.0 GB0.5 GB

Tip: go 8-bit.FP8 / INT8 is within about a percent of full precision in quality, takes half the memory and runs faster on RTX 40/50 and H100. Full precision only pays off for fine-tuning.

What to get for this model

Massed ComputeOpen Massed Compute

Your own GPU: FP8 / INT8 — from RTX 2060 · F32 — from RTX 2060

Massed Compute hourly prices, November 2026. The discounted price applies with the coupon at checkout.

Estimate: weights plus working memory. Long videos, high resolutions and long contexts need more.

Run it on Massed Compute

The script downloads the model to the rented machine. How to run it is in the model card.

  1. 1. Rent a machine with a fitting GPU

    Open Massed Compute
  2. 2. Paste the script into the machine's terminal

    Connect over SSH (the Massed Compute panel shows the command) or open a terminal on the machine's desktop.

    HF_TOKEN / CIVITAI_TOKENThe script starts with token lines: paste yours between the quotes. The Hugging Face token is for gated models, the Civitai one for Civitai models.

    # rimur01/diffusion_so101_assignment4_50k
    mkdir -p ~/models
    curl -LsSf https://astral.sh/uv/install.sh | sh
    source "$HOME/.local/bin/env"
    uv tool install "huggingface_hub[cli]"
    hf download rimur01/diffusion_so101_assignment4_50k --local-dir ~/models/diffusion_so101_assignment4_50k
    # The model is in ~/models. How to run it is in the model card (link above).
  3. 3. Use it

    The model is in ~/models. How to run it is in the model card (link above).

Stop the machine in the panel when you are done: it bills by the minute while it runs.

How to speed it up

Picked for this model type. Choose your hardware above to see what fits you exactly.

  • FlashAttention 44.0.0b40from November 18, 2026new release

    fastest exact attention on B200

    Kernels for Blackwell (beta). Exact attention without quantisation — fine for everything, LLMs included.

    How to enable: pip install flash-attn-4; vLLM and SGLang pick it up themselves.

    Earlier generations (2)
    • FlashAttention 3 — The Hopper (H100/H200) version with FP8.
    • FlashAttention 2 — The standard for RTX 30/40/50 and A100: exact attention without the memory cost of long contexts.
  • torch.compile2.15.0from October 28, 2026new release

    10–30% for free

    Compiles the model for your GPU. Slower first run, faster afterwards.

    How to enable: model = torch.compile(model); in ComfyUI the TorchCompileModel node.

Model family

New posts about versions of diffusion_so101_assignment4_50k (quants, finetunes, LoRAs) will show up under the bell.

No known relatives of this model yet. All derivatives on Hugging Face

About the model

rimur01/diffusion_so101_assignment4_50k
Author
rimur01
Category
Robotics
Released
Posted

On NeyroLearn

Hugging Face

220

Ask AI about this model

Answered by Gemini 4 · the chat is kept only in this tab

Comments