任务池

共 11 条，当前显示第 1–11 条（第 1/1 页）

• Performance degrades over time when running in Docker with Nvidia GPU

  ollama/ollama·#4846·Go·★ 181359·457 天未动·12 条评论

  bugperformancenvidia+1

  80

  可认领
• First value different on CUDA/ROCM when setting `seed`

  ollama/ollama·#4990·Go·★ 181359·584 天未动·2 条评论

  bugnvidiaamd

  80

  可认领
• gemma2 27b is too slow

  ollama/ollama·#5536·Go·★ 181359·704 天未动·4 条评论

  bugperformancenvidia+1

  79

  可认领
• Support split-mode=row for nvlink optimization

  ollama/ollama·#6131·Go·★ 181359·779 天未动·1 条评论

  feature requestnvidia

  72

  可认领
• Long context models don't split memory correctly leads to OOM error

  ollama/ollama·#4212·Go·★ 181359·865 天未动·0 条评论

  bugnvidiagpu

  61

  可认领
• Only Detecting One MIG Instance

  ollama/ollama·#4814·Go·★ 181359·557 天未动·6 条评论

  bugnvidia

  58

  可认领
• Unsupported nVidia cards still cause Cuda Toolkit/Drivers to be installed

  ollama/ollama·#674·Go·★ 181359·745 天未动·0 条评论

  bugnvidia

  57

  可认领
• CUBLAS_STATUS_ALLOC_FAILED with deepseek-coder-v2:16b

  ollama/ollama·#5395·Go·★ 181359·602 天未动·11 条评论

  bugnvidiamemory

  54

  可认领
• `Ollama` fails to work with `CUDA` after `Linux` suspend/resume, unlike other `CUDA` services

  ollama/ollama·#5464·Go·★ 181359·450 天未动·19 条评论

  bugnvidia

  51

  可认领
• Ask user to restart Ollama after Nvidia driver updates

  ollama/ollama·#4396·Go·★ 181359·852 天未动·0 条评论

  feature requestnvidiagpu

  49

  可认领
• Ollama is not using the 100% of RTX4000 VRAM (18 of 20GB)

  ollama/ollama·#3078·Go·★ 181359·620 天未动·29 条评论

  performancenvidia

  33

  可认领