Skip to content
Robotics

smolvla-so101-piano-flow

model for controlling a piano-playing robot

2 min readupdated

This compact vision-language-action system (a multimodal model that combines video, text, and physical actions) was trained with the LeRobot library on top of lerobot/smolvla_base. Built for the so_follower robot, it controls physical movements using feeds from three cameras and joint state data. In practice, it recognizes and presses specific keys on a musical instrument based on target notes. It is useful for robotics researchers and developers working with physical hardware on LeRobot.

Key features

  • Compact SmolVLA architecture designed to run on consumer hardware.
  • Combines data from three cameras (front, ears) for spatial orientation.
  • Trained on a 300-episode dataset of piano key-pressing tasks.
  • Full support for inference and fine-tuning using LeRobot ecosystem tools.

What you can do

  • Run the control policy on a physical so_follower robotic arm to play keys.
  • Fine-tune the model on your own motion datasets using LeRobot utilities.
  • Visualize the training dataset through the project's official Spaces.

Example

Running the pre-trained control policy to press a target key:

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=PawelLach/train_2026-11-17_22-11-36 \
  --task="play re, the second key" \
  --duration=60
Read moreDetails
Base model: lerobot/smolvla_base (0.45B parameters)
License: apache-2.0
Training: 10,000 steps, batch size 64, adamw optimizer, LR 0.0001
Dataset: 300 episodes, 25,308 frames at 30 FPS
Discuss Telegram 0

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 0.8 GB · ready files1.4 GB
  • FP8 / INT8recommendedweights 0.4 GB1.0 GB
  • 4-bitweights 0.2 GB0.8 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 · BF16 — 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.

    # PawelLach/smolvla-so101-piano-flow
    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 PawelLach/smolvla-so101-piano-flow --local-dir ~/models/smolvla-so101-piano-flow
    # 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

  1. SmolLM2-360Mfamily root
  2. SmolLM2-360M-Instruct
  3. SmolVLM-500M-Instruct
  4. SmolVLM2-500M-Video-Instruct
  5. smolvla_base
  6. smolvla-so101-piano-flow

New posts about versions of SmolLM2-360M (quants, finetunes, LoRAs) will show up under the bell.

Quantised versions 12All on Hugging Face
Fine-tunes 12All on Hugging Face
LoRAs and adapters 8All on Hugging Face
Merges 1All on Hugging Face

About the model

PawelLach/smolvla-so101-piano-flow
Author
PawelLach
Category
Robotics
Released
Posted

On NeyroLearn

Hugging Face

00

Ask AI about this model

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

Comments

Robotics

LoopedWAM

LoopedWAM