Track 10: Training LLMs, Evaluate, Compress and Ship
Evaluate Your Fine-Tune Properly
Produce a before/after table on a held-out set.
Part of Train and Fine-Tune Models (40 of 45)
Get notified on YouTube Udemy certificate: coming soon
Course
Fine-Tune Your Own LLM (Free GPU, Step by Step)
Udemy: coming soon
Tools
- DeepEval (open-source, freemium)
Watch first
Then try
- Make It Better: Data, Hyperparameters, Continued Pre-training, Multi-GPU
- Capstone: Fine-Tune for a Real Work Task
In this module
- Benchmarks Explained, and lm-evaluation-harness
- Evaluate Your Fine-Tune Properly
- Quantization: GGUF, GPTQ, AWQ, bitsandbytes
- Merge, Export, Run Your Model (and Model Merging)
- Distillation: Teach a Small Model From a Big One
- Capstone: Fine-Tune for a Real Work Task
- Model Merging With mergekit: SLERP, TIES, DARE
The course this lesson belongs to