REGISTERED TAG
Local AI course
Read mechanisms, local inference and adaptation in order.
01. Map AI, machine learning and LLMs
Match the problem you want to solve to a mechanism, rather than treating AI as one uniform tool.
02. Understand models, tokens and inference
Follow the transformation from text to numbers, then from numbers to the next token.
03. Learn the basics of training, loss and evaluation
Treat inference and training as different computations, and design evaluation before training.
04. Estimate memory for 24GB and 32GB Macs
Budget weights, KV cache, working memory and the operating system separately.
05. Read Hugging Face pages and model files
Check provenance, format and license so the model choice can be reproduced.
06. Run a small LLM locally
First observe input, output and resource use, before judging answer quality.
07. Choose a GPU platform and cost model
Align runtime, storage, stopping and data-transfer assumptions before comparing prices.
08. RAG: retrieve knowledge instead of packing it into weights
Separate retrieval from generation for questions that need changing facts and traceable sources.
09. Change behavior with SFT, LoRA and QLoRA
Organize additional training by its purpose and its implementation method.
10. Choose continued pretraining and preference optimization
Identify the goal and explain why a more demanding training method is needed.
11. Adapt to Japanese: vocabulary, notation and tasks
Evaluate Japanese language ability separately from domain expertise.
12. Handle medical terminology: sources, negation, units and human review
Distinguish terminology lookup and document processing prototypes from validated diagnostic use.
13. Connect local AI to images and 3D
Apply a shared view of computing resources while understanding different production pipelines.
14. Record experiments and integrate AI into a service
Turn small experiments into reproducible decisions and controlled operation.