| Has anyone got this same CUDA error: an illegal memory access was encountered while using this model in vllm:v0.25.0? | lzachary | open | 0 | disc |
| Help!!!CUDA driver error: device not ready | Rnikng | closed | 0 | disc |
| vLLM 0.23.1rc1.dev774+g2b802ad5b.d20260706. . load this model encounter OOM !! | xuefeizzzzzz | open | 0 | disc |
| SGLang support? | danielus | open | 0 | disc |
| [need help] vllm 0.24.0 on dgx spark warn Your GPU does not have native support for FP4 | imshenshen | open | 0 | disc |
| doubled TPS NVFP4 vs FP8 | sSeeMyRolexXx | open | 0 | disc |
| Fallback to marlin kernel give wired/incorrect and sometime gable result | sapjunior | open | 0 | disc |
| this model hallucinates easily | batsclamp | open | 0 | disc |
| vLLM + HGX B200: Your GPU does not have native support for FP4 computation but FP4 quantization is being used. | cgelias | open | 0 | disc |
| Works via Ollama, but the ~40 tok/s MTP speedup requires raw llama-server (Ollama has no --spec-type flag) - plus vLLM lockup findings on DGX Spark | darkmatter2222 | open | 0 | disc |
| 为什么全是！！！！！！！！！ | jiecong | open | 0 | disc |
| Why does this say 18B? | darkmatter2222 | open | 0 | disc |
| NVFP4 vs FP8 on an RTX PRO 6000 Blackwell (vLLM 0.24): sadly, FP4 is still not faster than FP8 - Qwen3.6‑27B | janreges3 | open | 0 | disc |
| CUDA error: an illegal memory access was encountered using vllm(sm120) | shakhizat | open | 0 | disc |
| Does it support MTP? | sharon8811 | open | 0 | disc |
| will it work on rtx 5090 32gb vram ? | arunsahu44 | open | 0 | disc |
| nvidia/Qwen3.6-27B-NVFP4 outputs repeated "!" / "d" tokens on Blackwell SM120 | richgua | open | 0 | disc |
| Could we include BF16 as well as FP8 in the benchmarks please? | haydonryan | open | 0 | disc |
| Qwen-AgentWorld-35B-A3B-NVFP4? | nagelanping | open | 0 | disc |
| Approved | jasionkajakub | open | 0 | disc |