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

Sft Demos

Sft Demos is a public repository of Jupyter Notebook examples for fine-tuning large language models, with a focus on short-form instruction following, plus documented Transformers inference code.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Sft Demos — 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 it when you want code examples to study or adapt for LLM instruction tuning, or when you need the README's documented path for preparing a model for text-generation inference.

Who it is for

  • Developers exploring notebook-based LLM fine-tuning examples.
  • Learners studying instruction tuning and language-model text generation.

Documented uses

  • Examine examples of fine-tuning LLMs for short-form instruction following.
  • Review training runs organized by base model in the repository's _peft area.
  • Prepare a Hugging Face model for prompt-based text generation with Transformers.
Before you choose

Requirements & limits

The repository is public and identifies an Apache 2.0 license.

  • The README focuses on short-form instruction following rather than presenting a general-purpose hosted fine-tuning service.
  • The provided inference setup uses Python, PyTorch and Hugging Face Transformers, and specifies a GPU-enabled cluster.
  • The README points to separate _peft training-run and _eval evaluation-run directories.

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 the documented text-generation example

Prepare the Python environment and load the README's example model for inference.

Before you start

  • Python 3.12 or later is identified in the repository badge.
  • A GPU-enabled cluster, as specified by the README.
  • A Jupyter Notebook environment for running the notebook-style code.

Steps

  1. Install transformers, accelerate and bitsandbytes using the README setup snippet.

  2. Download dfurman/CalmeRys-78B-Orpo-v0.1 with the Hugging Face CLI command shown.

  3. Load the model tokenizer and create a Transformers text-generation pipeline using the documented PyTorch and device settings.

Expected outcome

The model and tokenizer are prepared in a text-generation pipeline for the README's prompt examples.

Guide limitations

  • This is an inference setup, not a documented fine-tuning run.
  • The README requires a GPU-enabled cluster; it does not specify a complete hardware sizing guide.
Documentation 1 ↗
Evidence and transparency

Sources & review

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