badmonster

7 ms·

91.
by badmonster 1y ago
ran into this again today ..
92.
by badmonster 1y ago
nice
93.
by badmonster 1y ago
The stark performance gap between current models and humans—especially the fact that even the top proprietary model only hits 40%—highlights how underexplored and underdeveloped multi-image spatial reasoning still is.
94.
MMSI-Bench: A Benchmark for Multi-Image Spatial Intelligence(arxiv.org)
2 points by badmonster 1y ago|1 comments
95.
by badmonster 1y ago
nice
96.
Pre-Training for Recommendation Unlearning(arxiv.org)
1 points by badmonster 1y ago|0 comments
97.
Build image search and query with natural language with vision model CLIP(cocoindex.io)
2 points by badmonster 1y ago|0 comments
98.
Self-Reflective Uncertainties: Do LLMs Know Their Internal Answer Distribution?(arxiv.org)
1 points by badmonster 1y ago|0 comments
99.
Interactive Post-Training for Vision-Language-Action Models(arxiv.org)
4 points by badmonster 1y ago|0 comments
100.
by badmonster 1y ago
impressive formalization effort that bridges deep number theory and formal methods
101.
A Formal Proof of Complexity Bounds on Diophantine Equations(arxiv.org)
93 points by badmonster 1y ago|11 comments
102.
Google Gemini LearnLM Update(blog.google)
1 points by badmonster 1y ago|0 comments
103.
Let Androids Dream of Electric Sheep: A Human-Like Image Implication Understand(arxiv.org)
1 points by badmonster 1y ago|0 comments
104.
SignGemma – a sign language understanding model(twitter.com)
3 points by badmonster 1y ago|1 comments
105.
by badmonster 1y ago
amazing project
106.
Will AI Tell Lies to Save Sick Children? Litmus-Testing AI Values Prioritization(arxiv.org)
2 points by badmonster 1y ago|0 comments
107.
by badmonster 1y ago
GRAT (Grouping First, Attending Smartly) is a training-free method that massively speeds up diffusion-based Transformers—achieving up to 35.8× faster generation for ultra-high-res images (8192×8192)—by leveraging the natural sparsity in att
108.
Training-Free Acceleration for Diffusion Transformers(arxiv.org)
3 points by badmonster 1y ago|1 comments
109.
Build with Jules, your asynchronous coding agent(blog.google)
1 points by badmonster 1y ago|1 comments
110.
Radiology explainer demo(huggingface.co)
3 points by badmonster 1y ago|0 comments
111.
Google I/O 2025: From research to reality(blog.google)
2 points by badmonster 1y ago|0 comments
112.
Build Real-Time Product Recommendation Engine with LLM and Graph Database(cocoindex.io)
2 points by badmonster 1y ago|0 comments
113.
by badmonster 1y ago
reinforcement learning can push LLMs beyond generation and into true performance optimization at the assembly level. achieving a 1.47x speedup over gcc -O3 is no small feat—especially considering -O3 is already highly optimized.
114.
by badmonster 1y ago
Today, Waymo operates a commercial robotaxi service in all of San Francisco as well as parts of the Peninsula. It also provides driverless rides to paying customers throughout several Silicon Valley cities, including parts of Mountain View,
115.
Waymo gets OK to expand robotaxi service into more of Silicon Valley(techcrunch.com)
7 points by badmonster 1y ago|1 comments
116.
Improving Assembly Code Performance with LLMss via Reinforcement Learning(arxiv.org)
13 points by badmonster 1y ago|2 comments
117.
by badmonster 1y ago
very cute!
118.
by badmonster 1y ago
btw congrats on the launch
119.
by badmonster 1y ago
How are custom collision effects like ‘explode’ or ‘gravity shift’ implemented on top of Matter.js’ core collision events?
120.
by badmonster 1y ago
wow this is a popular project^_^ Which GitHub Actions context variables are emulated or customizable in act, like github.event, github.actor, or secrets

More ›