.mobo
+1 i have the same behaviour with gpt-oss + the last release of vllm + litellm on 2xH100. I have 128k context.
hi, same problem here. I try to enable text-to-speech with the docker container but there is no "speaches-cli" in the "/home/ubuntu/speaches/.venv/bin" folder. I think we should be able to "docker...
It seems to be related to this issue : https://github.com/speaches-ai/speaches/issues/387
Instead of Kokoro, you can try Piper. Ran this in the docker container : ```huggingface-cli download rhasspy/piper-voices --include 'en/**/*' 'voices.json'```
Nice thanks !
Ok, here is a Dockerfile that runs on OpenShift / Kubernetes / ArgoCD ``` ARG BASE_IMAGE=nvidia/cuda:12.6.3-cudnn-runtime-ubuntu24.04 FROM ${BASE_IMAGE} LABEL org.opencontainers.image.source="https://github.com/speaches-ai/speaches" LABEL org.opencontainers.image.licenses="MIT" RUN apt-get update && \ DEBIAN_FRONTEND=noninteractive apt-get install...
+1 for reranker
amazing, please integrate this devs :)
thanks, +1 for the fix here