llm-fine-tuning

An innovative library for efficient LLM inference via low-bit quantization

Collection of resources for finetuning Large Language Models (LLMs).

synthetic dataset generation workflow using local file resources for finetuning llms.

Fine tune LLM with HuggingFace

Sustain-LC is a benchmarking environment for traditional and reinforcement learning based controls as well as LLM based control

Distributed Reinforcement Learning for LLM Fine-Tuning with multi-GPU utilization

The Personal Knowledge Graph You Didn’t Know You Already Wrote

A curated collection of evolution strategies and zeroth-order optimization for large language models.

This repository contains code associated with Neuro-LIFT: A Neuromorphic, LLM-based Interactive Framework for Autonomous Drone FlighT at the Edge

A sacred space for heartfelt conversations, where wisdom flows freely and memories gently fade like whispers at sunset.

CLI tool for generating high-quality synthetic datasets for LLM fine-tuning.

Análise Avançada de Dados com Causalidade e Aprendizado por Reforço

Fine-tune Qwen3.8-27B to validate Form I-9 compliance (M-274/8 CFR 274a.2), citing violations. Synthetic-only data.

Source-tracked catalog of abliterated, low-refusal, and community-reported open-weight and hosted language models

The course teaches how to fine-tune LLMs using Group Relative Policy Optimization (GRPO)—a reinforcement learning method that improves model reasoning with minimal data. Learn RFT concepts, reward design, LLM-as-a-judge evaluation, and deploy jobs on the Predibase platform.

Experiments in Latin dactylic hexameter generation with transformers: A hybrid post hoc feedback framework

No-code desktop app for generating high-quality synthetic datasets to fine-tune LLMs — plan-then-execute pipeline, LLM-as-judge, HuggingFace upload.

FlowerTune LLM on Coding Dataset

Unified Amazon SageMaker AI plugin for Codex, Kiro, and Claude Code — SDK v3, training, inference, HyperPod, monitoring, and warm pools.

Flipper Zero Sub-GHz RF dataset (280-1100 MHz, 9 countries) with 1500 Q&A pairs for LLM fine-tuning, fact-checked allocations, and a GPU-accelerated validation pipeline (Ollama Qwen 32B + DeBERTa NLI).