GitHub - okuvshynov/slowllama: Finetune llama2-70b and codellama on MacBook Air without quantization
Mistral-finetune
mistral-finetune is a light-weight codebase that enables memory-efficient and performant finetuning of Mistral's models. It is based on LoRA, a training paradigm where most weights are frozen and only 1-2% additional weights in the form of low-rank matrix perturbations are trained.
For maximum efficiency it is recommended to use a A... See more
mistral-finetune is a light-weight codebase that enables memory-efficient and performant finetuning of Mistral's models. It is based on LoRA, a training paradigm where most weights are frozen and only 1-2% additional weights in the form of low-rank matrix perturbations are trained.
For maximum efficiency it is recommended to use a A... See more
GitHub - mistralai/mistral-finetune
🔍 If you're GPU-poor like me, here's how I fine-tuned a 7B LLM using only cloud services for less than $125, eliminating the need for expensive and specialized hardware. I'm sharing a practical, cost-effective approach to LLM training with the tools that I used 🧵
Shaurya Rohatgix.comRecorded a video discussing the fundamental topics of LLM finetuning.
- Different LLM Fine-tuning approaches:
- Reinforcement Learning from Human Feedback (RLHF)
- Challenges of fine-tuning an LLM:
- Fundamentals of Low-Rank Adaptatio... See more
Rohan Paulx.com