Articles
Context length, throughput and latency: measuring generation
Four numbers describe a generation; none of them means anything without its conditions.
Model behaviour
Directly assigned
Articles
GGUF and its runtimes: one file, several engines
GGUF is a container; which engine reads it, and how, is a separate question.
Articles
Inference on CPU and GPU: where the work actually goes
The same model runs on a CPU, a GPU or both; what changes is where the work is done.
Articles
Model formats and compatibility: what a weights file has to declare
A weights file is not the model: format, architecture and tokenizer must all line up.
Articles
Quantization: how much precision a model can lose
Fewer bits per weight save memory; what that costs in quality must be measured.
Articles
RAM and VRAM: what has to fit before a model runs
Weights, KV cache and working memory compete for the same pool; fit is arithmetic first.
Articles
The context window and its practical limits
A shared token budget, not durable memory: what the window is and where it runs out.
Articles
Hallucination, uncertainty and verification
Why a fluent model answer can be false, and what checking it actually requires.
Articles
Temperature, top-p and output limits
How one token is chosen out of a distribution, and what a token limit actually bounds.