dgslibisey/MuSiQue
Viewer
•
Updated
•
22.4k
•
1.41k
•
21
Wei Wu
Wei-Wu
AI & ML interests
None yet
Recent Activity
liked
a dataset
15 days ago
dgslibisey/MuSiQue
liked
a model
21 days ago
jiulaikankan/Qwen2.5-14B-ReasonGenRM
commented on
a paper
about 2 months ago
Skywork-Reward-V2: Scaling Preference Data Curation via Human-AI Synergy
Organizations
liked
a
dataset
15 days ago
liked
a
model
21 days ago
commented
a
paper
about 2 months ago
reacted to
codelion's
post with 🔥
about 2 months ago
Post
2380
🚀 Just published: "OpenEvolve: Open-Source Evolutionary Code Optimization with Real-World GPU Kernel Discovery"
We built the first open-source implementation of Google's AlphaEvolve system and used it to automatically discover GPU kernel optimizations that outperform human engineers!
Key results:
- 21.8% average decode speed improvement on Apple Silicon
- 36.7% improvement on long-context transformer attention
- Discovered novel vectorization patterns and 2-pass softmax algorithm
The system evolved a Metal kernel for Qwen3's Grouped Query Attention from a basic 3-pass implementation into something with sophisticated Apple Silicon optimizations that would take experts months to discover manually. The evolved kernel automatically found the optimal vec<T,8> operations for 128-dim attention heads and fused softmax computation with value accumulation.
Really excited about the potential here - imagine evolutionary algorithms automatically discovering optimizations across all our AI infrastructure. What would you want to optimize with this approach?
Full write-up: https://huggingface.co/blog/codelion/openevolve-gpu-kernel-discovery
GitHub: https://github.com/codelion/openevolve
#AI #MachineLearning #GPU #OpenSource #Evolution #CodeOptimization #TransformerOptimization
We built the first open-source implementation of Google's AlphaEvolve system and used it to automatically discover GPU kernel optimizations that outperform human engineers!
Key results:
- 21.8% average decode speed improvement on Apple Silicon
- 36.7% improvement on long-context transformer attention
- Discovered novel vectorization patterns and 2-pass softmax algorithm
The system evolved a Metal kernel for Qwen3's Grouped Query Attention from a basic 3-pass implementation into something with sophisticated Apple Silicon optimizations that would take experts months to discover manually. The evolved kernel automatically found the optimal vec<T,8> operations for 128-dim attention heads and fused softmax computation with value accumulation.
Really excited about the potential here - imagine evolutionary algorithms automatically discovering optimizations across all our AI infrastructure. What would you want to optimize with this approach?
Full write-up: https://huggingface.co/blog/codelion/openevolve-gpu-kernel-discovery
GitHub: https://github.com/codelion/openevolve
#AI #MachineLearning #GPU #OpenSource #Evolution #CodeOptimization #TransformerOptimization
- 1 reply
liked
a
model
about 2 months ago
liked
2
models
5 months ago
liked
a
model
6 months ago
liked
3
models
7 months ago
liked
a
dataset
7 months ago
liked
a
model
8 months ago
liked
a
model
9 months ago
liked
a
Space
9 months ago
upvoted
a
paper
11 months ago
liked
2
datasets
12 months ago
liked
2
models
about 1 year ago
