ST-Pruner
Official implementation of "Don't Ask the LLM: Visual Token Pruning with Spectral Structure and Semantic Relevance" (WACV 2027)
Get picks like this daily. Today's launch numbers, the top launches, AI and tech news, and a daily opportunity pick, straight to your inbox.
GITHUB · launched Oct 22, 2026 · 20 upvotes (top 64%) · 1 view
Category and tags
- Horizontal
- Model & infra
- Developer
- AI-native
- Hobby / open-source project
- efficient-inference
- token-pruning
- vision-language-model
- wacv2027
Visit website ↗ Source listing ↗
Could you build this?
No This is the official code release for a WACV research paper on visual token pruning. Its value is the novel spectral and semantic pruning method, which requires research expertise to reproduce and extend.
What it would actually take: Building a real version means implementing the paper's spectral-structure and semantic-relevance pruning on a vision-language model's visual tokens, then matching its benchmark results. That needs deep ML research background, GPU experiments, and access to the same benchmark datasets and evaluation setup.
Competitors
Other products that read as similar to this one : 2579 launches clear the similarity bar, closest 8 shown.
Attention rank: #1545 of 2558 (itself plus its competitors, highest first, normalized so YC and Product Hunt are compared fairly).
Launched 357 days after the earliest competitor.
- ml-lensvlm · github · 2026-09-22 · 75 upvotes · similarity 0.58
- reki · github · 2026-10-15 · 24 upvotes · similarity 0.56
- KohakuGBC · github · 2026-10-22 · 33 upvotes · similarity 0.56
- genpark-locality-sensitive-hashing-lsh-cosine-index-skill · github · 2026-09-29 · 7 upvotes · similarity 0.56
- Tokensift, an open-sourced token-efficiency linter for LLM prompts · hn · 2026-08-29 · 6 upvotes · similarity 0.55
Show 3 more
- FlexGlue · github · 2026-10-25 · 73 upvotes · similarity 0.55
- Stanford-CS336 · github · 2026-10-03 · 15 upvotes · similarity 0.54
- zero2vllm · github · 2026-10-12 · 14 upvotes · similarity 0.54