We were about to shut down a machine learning project. Then we decided to take another look at the data...
High-quality data, effective preprocessing, and model optimization are essential for successful implementation of generative AI applications.
Some hard-earned lessons from the trenches shipping machine learning products.
Think of business goals and interventions before determining your ML metrics.
Start with the low-hanging fruit to reduce friction and build trust in your organization in adopting machine learning across operations.
Standard retrieval can only get you so far. Alignment, contextual retrieval, and reranking can improve your RAG pipeline considerably.
To systematically analyze and fix LLM errors, think of the process in terms of classic ML error analysis.
Experiment with all kinds of generative AI and LLM tools. You'll find out where the industry is headed and how you can prepare for it.
MonsterGPT is a ChatGPT assistant that creates fine-tuned LLMs (Mistral, Llama-2, Llama-3, Gemma, OpenELM, etc.) without requiring any coding.
Claude-llm-trainer is a Google Colab project that fine-tunes Llama-2-7B with a single description about the downstream task.