OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Developer Tools & SDKs

Torchdistill

Torchdistill is a modular, configuration-driven PyTorch framework for knowledge distillation and reproducible deep-learning experiments. It lets users define experiment components in declarative YAML, use implemented distillation methods, and access configurations, trained models and logs for research and benchmarking.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Torchdistill — OpenViST project profile in AI Developer Tools & SDKs.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Torchdistill when a PyTorch study benefits from configuration-led experiment definitions, reproducible distillation workflows, or inspection of intermediate representations.

Who it is for

  • Researchers reproducing or benchmarking knowledge-distillation experiments.
  • PyTorch developers who want to define training studies through YAML configuration.
  • Practitioners analyzing intermediate representations in teacher or student models.

Documented uses

  • Specify a deep-learning experiment’s components in a declarative YAML file.
  • Reproduce or compare implemented knowledge-distillation methods using available configurations and artifacts.
  • Capture selected intermediate model inputs or outputs without changing the model’s forward interface.
  • Set up general deep-learning training without a teacher by excluding teacher entries from configuration.
Before you choose

Requirements & limits

The README lists Python 3.9 or later and installation through pip; pipenv is also documented as an optional installation route. The repository is public and carries an MIT license.

  • YAML configuration defines components such as models, datasets, optimizers and losses.
  • The README reports 26 implemented knowledge-distillation methods.
  • ForwardHookManager can extract selected module inputs and outputs using module paths.
  • Teacher-free training is supported by omitting teacher entries from the configuration.
  • Python 3.9 or later is required by the listed setup instructions.

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

Install Torchdistill with pip

Install the Torchdistill package in a Python environment.

Before you start

  • Python 3.9 or later
  • pip

Steps

  1. Run the command below.

    Documented command
    pip3 install torchdistill

  2. Use the installed package in your Python environment.

Expected outcome

Torchdistill is installed from the package index for use in PyTorch experiments.

Guide limitations

  • This covers package installation only; it does not configure or run an experiment.
  • The repository also documents pipenv and repository-based installation as alternatives.
Documentation 1 ↗
Evidence and transparency

Sources & review

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