Skip to content
Robotics

BBBXR/raids-hkpc-checkpoints

robotic policy checkpoints for LeRobot

1 min readupdated

A collection of weight sets and configurations for robot control, trained on the pi0.5 architecture (from Physical Intelligence) and ACT. The repository includes ready-to-use full inference parameters, delta-weights for step reconstruction, and historical training checkpoints. The package is intended for researchers and engineers working with physical robotics and the LeRobot platform.

Key features

  • pi0.5 models: Includes full inference parameters for stages 1-17 and 14 independent delta-weights for base parameters.
  • ACT models: ACT policy checkpoints (including Stage 9.5 and Stage 17) with task bindings and normalization.
  • Tokenizer: Ships with the PaliGemma tokenizer for working with pi0.5.
  • Historical data: Includes early training checkpoints and normalization configurations.

What you can do

  • Policy testing: Apply pre-trained pi0.5 and ACT weights to evaluate autonomous robot behavior.
  • Delta assembly: Reconstruct selected training steps using an exporter and delta-weights.
  • LeRobot integration: Use checkpoints in the LeRobot ecosystem for further experiments and deployment.

Example

Use the configuration and weight files from the repository to load control policies through the LeRobot infrastructure.

Read moreDetails
Library: lerobot
Task: robotics
Base model: pi0.5 (Physical Intelligence) / ACT
Limitations: Gemma license requirements and redistribution rules.
Discuss Telegram 1

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 0.5 GB · ready files1.0 GB
  • FP8 / INT8recommendedweights 0.2 GB0.8 GB
  • 4-bitweights 0.1 GB0.7 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.

    # BBBXR/raids-hkpc-checkpoints
    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 BBBXR/raids-hkpc-checkpoints --local-dir ~/models/raids-hkpc-checkpoints
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to raids-hkpc-checkpoints
  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 raids-hkpc-checkpoints (quants, finetunes, LoRAs) will show up under the bell.

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

About the model

BBBXR/raids-hkpc-checkpoints
Author
BBBXR
Category
Robotics
Released
Posted

On NeyroLearn

Hugging Face

01

Ask AI about this model

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

Comments