Skip to content
Other

SPEED+ Pose Estimator: Checkpoints for Spacecraft Pose Estimation Research

2 min readupdated

The model is designed to estimate the 6DoF (six degrees of freedom) pose of the Tango spacecraft from a single image. The architecture combines a ResNet-50 backbone and a fully connected head (MLP head), trained exclusively on synthetic data from the SPEED+ benchmark. The weight set includes 17 experimental checkpoints with different seeds and augmentation options for studying transfer from simulator to real domains (Lightbox and Sunlamp). The tool targets researchers studying model robustness and the simulator misspecification problem.

Key features

  • Direct 6DoF pose regression based on ResNet-50 and an MLP head.
  • Training exclusively on the synthetic SPEED+ dataset (28,000 images).
  • 17 experimental checkpoints with various seeds and augmentation settings.
  • Result reproducibility for analyzing the sim-to-real gap.

What you can do

  • Run the results table reproduction script to verify the research.
  • Test model inference on images of the Tango spacecraft.
  • Compare the accuracy of models with a frozen backbone and a trainable layer 4.

Example

import torch
from huggingface_hub import hf_hub_download
from PIL import Image
from speedplus_pose.model import SpacecraftPoseEstimator
from speedplus_pose.normalisation import TranslationNormaliser
from speedplus_pose.transforms import build_eval_transform

repo = "lkid08/spacecraft-pose-estimation"
hf_hub_download(repo, "config.json")
path = hf_hub_download(repo, "unfrozen_aug_seed42.pt")
checkpoint = torch.load(path, weights_only=False, map_location="cpu")

model = SpacecraftPoseEstimator()
model.load_state_dict(checkpoint["model_state_dict"])
model.eval()

normaliser = TranslationNormaliser(checkpoint["train_means"], checkpoint["train_stds"])
image = Image.open("img000001.jpg")
tensor = build_eval_transform()(image)

with torch.no_grad():
    output = model(tensor.unsqueeze(0))

translation = normaliser.remove(output[:, :3])
quaternion = output[:, 3:]
quaternion = quaternion / quaternion.norm(dim=1, keepdim=True)

print(translation, quaternion)
Read moreDetails
Benchmarks (orientation error E_R): layer 4 backbone with augmentation shows an error of 0.987 rad on Lightbox.
Hardware requirements: training was performed in Google Colab and on Apple M5.
Limitations: the model is not intended for real-world spaceflight use, accuracy on real data drops, and model confidence remains high during erroneous predictions.
Training details: input size 320x512 in grayscale, target object is the Tango spacecraft.
Discuss Telegram 1 18

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 0.2 GB · ready files0.7 GB
  • FP8 / INT8recommendedweights 0.1 GB0.6 GB
  • 4-bitweights 0.1 GB0.6 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.

    # lkid08/spacecraft-pose-estimation
    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 lkid08/spacecraft-pose-estimation --local-dir ~/models/spacecraft-pose-estimation
    # 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 spacecraft-pose-estimation (quants, finetunes, LoRAs) will show up under the bell.

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

About the model

lkid08/spacecraft-pose-estimation
Author
lkid08
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

181

Ask AI about this model

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

Comments