Learn to accelerate Supervised Fine-Tuning (SFT) and Direct Preference Optimization (DPO) training for Large Language Models through a detailed video tutorial that explores two free Jupyter notebooks. Dive into practical implementations using HuggingFace-compatible scripts for training LLama or Mistral models, with step-by-step demonstrations of the free version's capabilities. Access comprehensive examples including Alpaca with Mistral 7b implementation and DPO Zephyr training, complete with direct links to ready-to-use Google Colab notebooks for hands-on experimentation in AI model training and optimization.
Overview
Syllabus
FASTER Code for SFT + DPO Training: UNSLOTH
Taught by
Discover AI
Related Courses
-
Generative AI Advance Fine-Tuning for LLMs
-
ORPO: A New Preference-Aligned Training Method for Large Language Models
-
Coding RLHF on LLama 2 with LoRA, 4-bit Quantization, TRL and DPO
-
From Research to Production: Fine-Tuning and Aligning LLMs
-
Combined Preference and Supervised Fine-Tuning with ORPO
-
Direct Preference Optimization - Fine-Tuning LLMs Without Reinforcement Learning
