Skip to content
Robotics

WeaveMem YAM 10K

Inference service for robotics models

2 min readupdated

The package contains three native π₀.₅ models trained on the YAM robotics dataset. The server accepts camera frames, robot state, and instructions via WebSocket, returning an array of 50 predicted actions for 14 degrees of freedom. Decisions are based on Orbax inference parameters and include a frozen OpenPI runtime, tokenizer, and CD-LAM encoder. The model is designed for researchers and engineers working on robotic manipulator control.

Key features

  • Support for three checkpoints with various sync and async modes (sync, kv, async_lag5).
  • Processing data from multiple cameras and a 14-dimensional robot state.
  • Built-in WebSocket server for fast transmission of commands and observations.
  • Use of the CD-LAM encoder for working with movement history.

What you can do

  • Deploy the inference server on Linux with an NVIDIA GPU and run robot control.
  • Set up a Python client to send telemetry and receive action arrays [50, 14].
  • Integrate asynchronous modes with a 5-frame lag for smooth real-time operation.

Example

from websockets.sync.client import connect
from openpi_client import msgpack_numpy

with connect("ws://127.0.0.1:8765", compression=None, max_size=None) as ws:
    metadata = msgpack_numpy.unpackb(ws.recv())
    packer = msgpack_numpy.Packer()
    call = lambda msg: (ws.send(packer.pack(msg)), msgpack_numpy.unpackb(ws.recv()))[1]
    call({"reset": True, "seed": 42})
Read moreDetails
Hardware requirements: Linux x86-64, Python 3.11, NVIDIA GPU with CUDA 12 support (tested on H100).
Communication protocol: binary MessagePack over WebSocket.
Action format: left arm (6 joints + gripper), right arm (6 joints + gripper).
Discuss Telegram 0

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 2.6 GB · ready files3.4 GB
  • FP8 / INT8recommendedweights 1.3 GB2.0 GB
  • 4-bitweights 0.7 GB1.3 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.

    # Tingxiang/weavemem-yam-10k
    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 Tingxiang/weavemem-yam-10k --local-dir ~/models/weavemem-yam-10k
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to weavemem-yam-10k
  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 weavemem-yam-10k (quants, finetunes, LoRAs) will show up under the bell.

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

About the model

Tingxiang/weavemem-yam-10k
Author
Tingxiang
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