Goblin
News
AI news, distilled.
|
News
About
News
About
Filtered by:
dflash
Clear
Titles
Summaries
6
Coding AI Inference Stack Achieves 3x+ Speedup with Trained Speculators and Custom GPU Kernels
Research
1
Jun 22
6
Coding AI Inference Stack Achieves 3x+ Speedup with Trained Speculators and Custom GPU Kernels
Research
· 1 src · Jun 22
Discuss
7
DFlash Speculative Decoding Achieves 4.3× LLM Throughput With Block Diffusion and KV Injection
Research
1
Jun 16
7
DFlash Speculative Decoding Achieves 4.3× LLM Throughput With Block Diffusion and KV Injection
Research
· 1 src · Jun 16
Discuss
Filters
Signal
Title
Category
Sources
Posted
Discuss
