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

Llm2Llm

Llm2Llm is open-source Python research code for reproducing the LLM2LLM paper’s iterative data-enhancement experiments, with documented instructions for GSM8K.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Llm2Llm — 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 Llm2Llm when you want to reproduce its documented GSM8K research workflow, from seed-data preparation through iteration-level reporting.

Who it is for

  • Researchers reproducing the documented GSM8K experiments
  • Developers studying iterative data enhancement and model performance across experiment iterations

Documented uses

  • Prepare seed data for the GSM8K experiment.
  • Run the experiment iterations using the supplied pipeline.
  • Generate a detailed per-iteration model performance report.
  • Retain the experiment data and model checkpoints produced by the workflow.
Before you choose

Requirements & limits

The repository is public and licensed under MIT.

  • The provided reproduction procedure is for GSM8K; the README does not include instructions for other datasets.
  • The workflow requires a LLaMA-2-7B copy, the appropriate dataset, seed-data preparation, and experiment configuration.

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

Reproduce the documented GSM8K experiment

Run the repository’s GSM8K experiment pipeline and produce iteration-level performance results.

Before you start

  • A copy of LLaMA-2-7B
  • The appropriate dataset and the GSM8K grade-school-math repository
  • The LLM2LLM repository and its experiment configuration

Steps

  1. Clone the GSM8K dataset repository in the GSM8K directory.

  2. Run generate_seed_data.py and adjust SUBSAMPLE_SPLIT to obtain seed data.

  3. Check the settings in GSM8K/config.yaml, then run the command below.

    Documented command
    python GSM8K/generator_data.py GSM8K/config.yaml

  4. In the experiment folder, run ./run_all.sh; after the iterations finish, run the documented report_results.py command with the test data path and experiment folder.

Expected outcome

The README says the workflow produces an output folder with data and model checkpoints, and a detailed breakdown of model performance at each iteration.

Guide limitations

  • The README provides reproduction instructions for GSM8K and says instructions for other datasets were planned.
  • The guide reflects the repository instructions; it is not a report of an execution test.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
16 October 2026
Original website title
SqueezeAILab/LLM2LLM · Official homepage title · retained in its source language
How we review sourcesSuggest a correction