franka-pico4-smolvla

Exile051112/franka-pico4-smolvla

franka-pico4-smolvla is a machine learning model from Exile051112. Use it for the machine learning task on the model card, and read the license before you ship it in a product.

This repository contains the shared SmolVLA base and metadata plus independent LoRA adapters trained on the Franka/Pico4 LeRobot datasets. Adapter folders are separate checkpoints; they are not mathematically merged.

Downloads · 30 days

Access

Public

Updated Aug 29, 2026

Repo size

1.3 GB

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors1.3 GB · 96%

At a glance

Access: Public
Created: Aug 28, 2026
Updated: Aug 29, 2026
SHA: b0f9918d

Model card

From the Hugging Face model README

Franka Pico4 SmolVLA Models

Folder	Training data	Status
`base`	Shared `lerobot/smolvla_base` weights	Shared base
`vlm_metadata`	Shared tokenizer/processor metadata	Shared support files
`c1_red_yellow_real`	5 red real + 5 yellow real	10k steps
`c2_red_yellow_blue_real`	5 red real + 5 yellow real + 5 blue real	10k steps
`c3_red_yellow_blue_edited`	5 red real + 5 yellow real + 5 blue edited	10k steps
`c4_red_yellow_real_blue_sim`	5 red real + 5 yellow real + 5 blue simulated	10k steps
`c5_red_yellow_blue_sim`	5 red simulated + 5 yellow simulated + 5 blue simulated	10k steps

All experiment folders contain checkpoints at 2k, 4k, 6k, 8k, and 10k steps. The adapters use PEFT LoRA (r=64, alpha=64) and require the shared base and metadata when loaded. The evaluation target is the blue Franka/Pico4 setup.

Training uses the official PyTorch LeRobot workflow. Dataset feature dimensions are the robot-specific 17-D state and 10-D TCP action with top and wrist RGB camera streams at 30 FPS.

Created: Aug 28, 2026
Updated: Aug 29, 2026