unsloth/studio/backend/core/inference
Daniel Han cab3ed41e9 fix: fetch top search result page content for web_search tool
The web_search tool previously returned only titles, URLs, and short
snippets from DuckDuckGo.  Small models would loop repeatedly trying
different search queries because the snippets never contained enough
data to answer the question.

Now fetches the actual page content for the top 1-2 search results
(best-effort, capped at 6000 chars and 10s timeout per fetch) and
includes it in the tool response.  This gives the model real data to
work with on the first search instead of endlessly retrying.

Uses only stdlib (urllib + re) so no new dependencies are needed.
2026-03-27 13:34:52 +00:00
..
__init__.py Final cleanup 2026-03-12 18:28:04 +00:00
audio_codecs.py studio: per-model inference defaults, GGUF slider fix, reasoning toggle (#4325) 2026-03-16 06:37:55 -07:00
defaults.py studio: web search, KV cache dtype, training progress, inference fixes 2026-03-17 00:30:01 -07:00
inference.py fix: system prompt ignored in unsloth inference (#4528) 2026-03-24 04:01:33 -07:00
llama_cpp.py [pre-commit.ci] auto fixes from pre-commit.com hooks 2026-03-27 13:33:58 +00:00
orchestrator.py feat(studio): infinite scroll for recommended models list (#4414) 2026-03-18 03:17:01 -07:00
tools.py fix: fetch top search result page content for web_search tool 2026-03-27 13:34:52 +00:00
worker.py Fix inference failing for transformers 5.x models (trust_remote_code) (#4652) 2026-03-27 04:51:30 -07:00