We unified the interfaces of instruction-tuning data (e.g., CoT data), multiple LLMs and parameter-efficient methods (e.g., lora, p-tuning) together for easy use. We welcome open-source enthusiasts to initiate any meaningful PR on this repo and integrate as many LLM related technologies as possible. 我们打造了方便研究人员上手和使用大模型等微调平台，我们欢迎开源爱好者发起任何有意义的pr！

No description available

Parameter-Efficient Sparsity Crafting From Dense to Mixture-of-Experts for Instruction Tuning on General Tasks (EMNLP'24)

CodeUp: A Multilingual Code Generation Llama-X Model with Parameter-Efficient Instruction-Tuning

Astraios: Parameter-Efficient Instruction Tuning Code Language Models

Parameter-Efficient Instruction Tuning of Large Language Models For Extreme Financial Numeral Labelling

In this work, we address the dual challenge of accuracy and interpretability by adapting a multimodal large language model to perform both precise PMA prediction and clinical-relevant explanation generation. We introduce a parameter-efficient fine-tuning strategy using Instruction Tuning and Low-Rank Adaptation applied to the Qwen2.5-VL-7B model.

End-to-End parameter-efficient domain adaptation and instruction fine-tuning of Qwen3-1.7b LLM

Appendices of paper "Stepwise Schema-Guided Prompting Framework with Parameter Efficient Instruction Tuning for Multimedia Event Extraction".

This repository includes code and experiments for fine-tuning the microsoft/Phi-3-mini-4k-instruct model using Parameter-Efficient Fine-Tuning (PEFT) with LoRA. It is focused on instruction-based tuning and is optimized for generating natural language product advertisements.

AIST5030 Mini-Project: Parameter-Efficient Instruction Tuning via Orthogonal Finetuning

Question-Guided Multi-Task Instruction Tuning for Parameter-Efficient Dialogue State Tracking

Parameter-efficient fine-tuning pipeline for domain-specific LLMs using QLoRA, synthetic data generation, and instruction tuning.

A COMPARATIVE STUDY OF PARAMETER-EFFICIENT FINE-TUNING FOR INSTRUCTION-FOLLOWING LARGE LANGUAGE MODELS

Using p-tuning (a form of parameter-efficient fine-tuning) to tune Pythia models on natural language instructions.

Parameter-Efficient Fine-Tuning of FLAN-T5 using LoRA and IA³ on the Databricks Dolly 15K dataset for instruction tuning.

parameter efficient fine tuning on flan t5 opensource llm and comparison between instruction fine tuning and evaluation using ROUGE Score

Geometric Reprojection Instruction Tuning (GRIT): A parameter-efficient fine-tuning framework that combines LoRA with curvature-aware optimization and neural reprojection for efficient adaptation of vision-language models

A fine-tuned GPT-2 Medium model using LoRA (Low-Rank Adaptation) on the Alpaca instruction dataset. This project demonstrates efficient parameter-efficient fine-tuning for instruction-following tasks.

Instruction fine-tuning a FLAN-T5-Base model using the SQuAD Dataset. Parameter efficient fine tuning the FLAN-T5-Base model using LoRA on SQuAD

Flan-T5 model for dialogue summarization, integrating instruction-based fine-tuning, parameter-efficient fine-tuning (PEFT), and Reinforcement Learning with Human Feedback (RLHF) for output detoxification.

Parameter-efficient instruction fine-tuning using LoRA to adapt OPT-350M for code generation. Trains <1% of parameters (1.7M/350M) on CodeAlpaca-20k with SACREBLEU evaluation. Demonstrates practical PEFT techniques for SLMs.

This repository contains a Jupyter Notebook for fine-tuning the TinyLlama-1.1B model using Parameter-Efficient Fine-Tuning (PEFT) techniques, specifically LoRA. The model is trained on the Alpaca dataset to improve its instruction-following capabilities.

Implemented parameter-efficient instruction tuning by applying LoRA to TinyLlama and Pythia on LIMA and benchmarking zero-shot performance on HellaSwag, then applied an improved LoRA variant and compared results.

This repository contains two self-contained fine-tuning workflows: Fine-tunning LLama 2: Parameter-efficient fine-tuning of Llama-2-7B-Chat using QLoRA and PEFT on a small instruction dataset. FIne-tunning using unsloth: A fast, memory-efficient fine-tuning of Phi-3 Mini (4k instruct, 4-bit) using Unsloth with LoRA.

Fine-Tuning LLaMA 2 with Hugging Face Transformers This repository contains code and configuration for fine-tuning Meta's LLaMA 2 model using the Hugging Face transformers and peft (Parameter-Efficient Fine-Tuning) libraries. It supports training on custom datasets for instruction tuning, Q&A, or domain-specific NLP tasks.

LLMs based on transformers excel at abstractive summarization, but complete fine-tuning is usually unfeasible with restricted computing resources. This study assesses a Gemma-2B instruction-tuned model on the XSum summarization task, contrasting a baseline (without fine-tuning) system with parameter-efficient fine-tuning utilizing QLoRA.

This project Instruction fine-tunes the Phi-2 language model using QLoRA to generate Amazon product names and descriptions. It demonstrates efficient parameter-efficient fine-tuning on custom text data using low-bit quantization and Hugging Face tools.

Implemented transfer learning, Parameter-Efficient Fine-Tuning (PEFT), and Low-Rank Adaptation (LoRA) on the LLaMA 3.2-3B-Instruct model to generate RISC-V assembly instructions for load and store operations.

Fine-tuning Google’s Gemma 2B model using KerasNLP and the Dolly-15k dataset in Google Colab. This project covers data preparation, LoRA-based parameter-efficient fine-tuning, and generation using instruction-context prompts. Focused on low-resource environments and practical GenAI training.