The H2O.ai team taught how to make and fine-tune your own LLM using open assistant datasets. Every organization needs to own its data, code, algorithms, and models. Every company needs to build and fine-tune its own GPT. At this meetup, participants learned:

• Choices in fine-tuning and different LLM backbones.
• Different quantization options Int8, float16, etc.
• Different methods like LORA, Adaptive Filtering.
• Available evaluation metrics and comparisons against GPT3.5

In attendance: