Qwen output breaks in Docker on Ubuntu, but works on Windows 11 with WSL 2
awitod · reddit · 2026-07-23
A Reddit user reports that a llama.cpp Docker image produces bad output for Qwen 3.6 35B and 27B on an Ubuntu host, while the exact same container works on Windows 11 via Docker Desktop and WSL 2.
They suspect the host-side ROCm stack: drivers and packages are up to date, and the mismatch only appears on Ubuntu, pointing to an issue in the ROCm environment rather than the image itself.
More from Infra
- Resin lets an agent build standalone LLM inference engines; vLLM and SGLang optimize DeepSeek kernels differently — AccBalanced · 2026-10-27
- Local benchmark: Flint matches Qwen3.6 quality at 50/52 while running 3.6x faster on M4 Pro — RichardSocher · 2026-10-27
- Anthropic S-1: CSP marketplace revenue jumped from 11% to 48% in three years — AccBalanced · 2026-10-27
- Alchemy adds effect/Workflow support across Durable Objects, Workflows and Lambda Durable Functions — samgoodwin89 · 2026-10-27
- Alchemy's Lambda Web Functions run plain Node web servers on Lambda, no OpenNext needed — samgoodwin89 · 2026-10-27
- Alchemy Cuts Cloudflare Bindings CPU Time by Up to 166× Via Profiling — samgoodwin89 · 2026-10-27