Open Models · first seen 02 November 2026, 07:28 UTC
qwenGPU runs Qwen3.5-2B with 45 compiled GPU layers
qwenGPU is a project that runs the Qwen3.5-2B model with 45 compiled GPU layers executing CUDA, aiming at efficient local inference.
qwenGPUOpenCodeOpen weights
- Params
- 2B
- Domain
- general
Read it at MarkTechPost →
Coverage: 3 sources, in order
-
MarkTechPost Read the article → -
MarkTechPost Read the article → -
MarkTechPost Read the article →
← Front page