A hands-on series for learning small language-model training from tokenizer to serving.
Latest article
Trace MiniMind's native PyTorch model, data pipeline, training stages, evaluation, checkpoints, API, and WebUI as explicit boundaries.
By publication date
01 → 09
Trace MiniMind's native PyTorch model, data pipeline, training stages, evaluation, checkpoints, API, and WebUI as explicit boundaries.
Compare MiniMind with fine-tuning a larger model, inference-only runtimes, hosted APIs, and educational notebooks by learning value, control, quality, cost, and operations.
A practical MiniMind deployment guide covering Python/PyTorch environments, single- and multi-GPU training, checkpoint management, API serving, and evaluation.
A source-backed MiniMind walkthrough for inference demos, native PyTorch training stages, OpenAI-compatible serving, and reproducible evaluation.
A practical capstone for versioned MiniMind datasets, stage receipts, evaluation cards, checkpoint lineage, safe serving, and evidence-backed experiments.
A source-backed MiniMind overview covering its 64M model, native PyTorch training stages, inference options, and reproducible learning path.
A reproducible MiniMind measurement plan for data throughput, GPU memory, training stages, benchmark quality, inference latency, and total experiment cost.
An operations guide for MiniMind data provenance, model artifacts, OpenAI-compatible serving, tool calls, secrets, monitoring, and incident recovery.
A fixture-driven method for reading MiniMind's tokenizer, native PyTorch forward pass, training loop, checkpoint resume, and OpenAI-compatible serving.
