Observed arrival · 2026-11-02
Xiaoli’s Blog: Small Experiments in Model Training
A Chinese-language technical blog documenting machine-learning experiments, from quantization and distributed training to fine-tuning and alignment.
○Open
⊠Login
$Paid
†Ads
✦Pretty
●Pro
◎Niche
◉Human
⚑Risk
ƒJS
- For
- Readers learning machine-learning systems and training methods
- Worth noticing
- The homepage reports a toy-model quantization test where int2 reaches zero legality while int4 retains a 1.90 perplexity.
Field notes
The archive groups posts into categories such as distributed training and deployment, fine-tuning and alignment, and mathematics-related subjects. Its summaries name setups as well as outcomes: one describes four-process Gloo tests on a laptop without a GPU, while another compares communication bucketing with sending tensors individually. The footer says articles use a CC BY-NC-SA 4.0 license and identifies Astro and the Fuwari theme.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
○OpenPublic substance visible
✦PrettyNotable craft visible
●ProPolished or operationally mature
◎NicheUnusually specific use
◉HumanPersonal, local, civic, or handmade
ƒJavaScriptBrowser-side code central
One card from the complete issue