Lit-GPT官網(wǎng)

快捷微調(diào)LLM

網(wǎng)站服務(wù)：AI開(kāi)源項(xiàng)目，ai，AI編程，AI開(kāi)源項(xiàng)目，ai。

Lit-GPT簡(jiǎn)介

Hackable implementation of state-of-the-art open-source LLMs based on nanoGPT. Supports flash attention， 4-bit and 8-bit quantization， LoRA and LLaMA-Adapter fine-tuning， pre-training. Apache 2.0-licensed. – GitHub – Lightning-AI/lit-gpt: Hackable implementation of state-of-the-art open-source LLMs based on nanoGPT. Supports flash attention， 4-bit and 8-bit quantization， LoRA and LLaMA-Adapter fine-tuning， pre-training. Apache 2.0-licensed.
一個(gè)基于 Lit-LLaMA 和 nanoGPT 實(shí)現(xiàn)，支持 LLaMa 2、Mistral 7B、Mixtral 8x7B 等模型的微調(diào)。支持閃存注意力，Int8和GPTQ 4bit量化，LoRA和LLaMA-Adapter微調(diào)，預(yù)訓(xùn)練等功能。

Lit-GPT官網(wǎng)入口網(wǎng)址

https://github.com/Lightning-AI/lit-gpt

OpenI小編發(fā)現(xiàn)Lit-GPT網(wǎng)站非常受用戶(hù)歡迎，請(qǐng)?jiān)L問(wèn)Lit-GPT網(wǎng)址入口試用。