🎯 Quantization

The items that name it 11

• Show HN: A language model that lives in a boot sector
  repo2026-10-31

  Weights are quantized to int8 with power-of-two scales so the boot sector can run the forward pass with imul and a single idiv.

• nemotron-halloween
  writeup2026-10-30

  The 4-bit quantized Bolt E4B release is the default response model to fit the memory and compute budget.

• Meta’s AI-distilled model detector catches 79% at a 0.1% false-positive rate
  article2026-10-30

  Compresses a model to reduce running costs. The WM-Basin detection signal survived it.

• [AINews] Building AIE CODE: The top Coding Agent Conference
  article2026-10-30

  FP8, NVFP4 and 1-bit GGUF checkpoints reduce Mistral Large 4 size; NyoomFloat12 compresses BF16 weights losslessly to 12 bits.

• EmbeddingGemma 2: an open, lightweight multimodal embedding model
  article2026-10-06
• OpenAI Banned PewDiePie Twice While He Built Ajax, an Uncensored AI That Will Run on Your PC
  article2026-10-03
• The Roadmap to Mastering LLM Inference Optimization
  article2026-09-21
• bilibili/Index-Translate
  repo2026-09-15It is written here as Quantization (GGUF, FP8 W8A8, NVFP4 W4A4).

  GGUF, FP8 W8A8, NVFP4 W4A4

• datawhalechina/zero-to-sglang
  course2026-08-27
• patchy631/time-to-first-token
  repo2026-08-02It is written here as Quantization (FP8, INT4/AWQ, GPTQ, SmoothQuant, LLM.int8()).

  FP8, INT4/AWQ, GPTQ, SmoothQuant, LLM.int8(

• Retrochainer Post
  thread2026-08-01

In every workspace you can read 12