南栖's repositories
character_AI_open
Generate multi-round conversation roleplay data based on self-instruct and evol-instruct.
deepspeed-grpo-qlora-vllm
This repository, deepspeed-grpo-qlora-vllm, provides a complete framework for fine-tuning LLMs using Group Relative Policy Optimization (GRPO) on 4-bit quantized models (QLoRA). It utilizes DeepSpeed ZeRO-3 for scalable training and integrates with a VLLM server to dynamically serve the fine-tuned LoRA adapters.
attention_sinks_autogptq
attention_sinks can use autogptq,and support all model at autogptq,like qwen baichuan,etc
Infinite-Evolution-AI
Iteratively Generating Complex Evolutionary Networked Instructions uncensored.
alignment-handbook
Robust recipes to align language models with human and AI preferences
alpaca_eval
An automatic evaluator for instruction-following language models. Human-validated, high-quality, cheap, and fast.
AutoGPTQ_cogvlm
An easy-to-use LLMs quantization package with user-friendly apis, based on GPTQ algorithm.
Bert-VITS2
vits2 backbone with bert
character_AI_open_evol
Achieve 2–3× roleplay performance through LLM self-iteration with MCTS and evol instruction.
Emotional-ai
Emotional ai
InfLLM
The code of our paper "InfLLM: Unveiling the Intrinsic Capacity of LLMs for Understanding Extremely Long Sequences with Training-Free Memory"
wxbot
PC微信Hook模块、Hook WeChat / 微信逆向、微信机器人、WeChatRobot
Evol_Mctsr
Achieve a 2x-3x performance improvement through LLM self-iteration with MCTS and evo instruction.
hallucination-leaderboard
Leaderboard Comparing LLM Performance at Producing Hallucinations when Summarizing Short Documents
MemAgent
A MemAgent framework that can be extrapolated to 3.5M, along with a training framework for RL training of any agent workflow.
mini-omni2
Towards Open-source GPT-4o with Vision, Speech and Duplex Capabilities。
nemori
A minimalist MVP demonstrating a simple yet profound insight: aligning AI memory with human episodic memory granularity. Shows how this single principle enables simple methods to rival complex memory frameworks for conversational tasks.
OpenAlita
Open Source Implementation of Alita: Generalist Agent Enabling Scalable Agentic Reasoning with Minimal Predefinition and Maximal Self-Evolution
OpenAlpha_Evolve
OpenAlpha_Evolve is an open-source Python framework inspired by the groundbreaking research on autonomous coding agents like DeepMind's AlphaEvolve.
theLMbook
This is the official repository for The Hundred-Page Language Models Book by Andriy Burkov
titans
Titans paper implementation
trl
Train transformer language models with reinforcement learning.
vllm-gptq
A high-throughput and memory-efficient inference and serving engine for LLMs