OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Models

Hlt

Hlt is a Python implementation of a transformer-based policy model for humanoid robotics. It combines a state-policy MLP/FFN with a transformer that processes observation and action tokens to produce the next action sequence, with documented inputs including video and instruction embeddings.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Hlt — OpenViST project profile in AI Models.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Hlt when exploring or prototyping a model that maps video and instruction embeddings to action-sequence outputs, rather than when seeking a documented end-to-end robot-control application.

Who it is for

  • Robotics researchers studying humanoid locomotion and action prediction
  • Developers prototyping transformer-based robot policy models
  • AI practitioners building research harnesses for action-sequence prediction

Documented uses

  • Inspect an implementation of a transformer model for humanoid robot locomotion.
  • Pass video and instruction embeddings through an HLT model to obtain action-sequence outputs.
  • Prototype a research workflow around observation and action tokens.
Before you choose

Requirements & limits

The public repository specifies an MIT license and documents installation through `pip install hlt-torch`.

  • The example requires video tensors and instruction embeddings, with the README specifying example shapes.
  • The README demonstrates a forward call using randomly generated tensors; it does not describe a hardware-control workflow.
  • The implementation is presented as a model component, not a complete robotics application.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Set up and run the documented tensor example

Install the package, create an HLT model, and pass example-shaped video and instruction tensors to its forward call.

Before you start

  • Python environment with PyTorch available
  • Install the package using pip install hlt-torch

Steps

  1. Install the package with the command below.

    Documented command
    pip install hlt-torch

  2. Import torch and HLT from hlt_torch.model.

  3. Construct HLT using the parameter values shown in the README example.

  4. Create video input shaped (1, 3, 16, 112, 112) and instruction embeddings shaped (1, 10, 512), then call model(video, instructions).

Expected outcome

The example produces an output tensor from the model call.

Guide limitations

  • The README uses randomly generated dummy inputs, not recorded robot observations or instruction embeddings.
  • The example does not document training, robot hardware integration, or using the output to control a robot.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
13 October 2026
Original website title
kyegomez/HLT · Official homepage title · retained in its source language
How we review sourcesSuggest a correction