Baruch Berger
Baruch Berger
I also ran into this, it also happens with crux-lmdb. AdoptOpenJDK seems to not have this issue, thanks @AndreasEdvardsson
> :shadow/ouputs could be moved here instead and I'm still debating whether this would be a better place to have them or the manifest.edn. Feedback welcome. I did think it...
I've been trying out Angular2 with Dart and been pretty happy so far, it's definitely "halfway usable", even though still tagged as alpha. Also I believe it has been said...
I'm seeing the same error after updating from I think some version of cider 25 in my ClojureScript projects. Tried both melpa stable and 20201020..1058 . I'm not sure if...
My mistake, had mismatched nrepl middleware in the projects. Thanks!
Also hit attrs not having a prototype, it breaks serialization of docs in my project.
What happens on conflicts? How does it compare to CRDT's?
Thank you, however this is not high priority for us anymore! I was able to use jxlCreateInstance, jxlProcessInput and jxlFlush to decode images after adjusting the wasm example and building...
[Bug]: Qwen2.5-VL-72B-Instruct-AWQ error with TP=2 and low throughput (~2 tokens/s) on VLLM_USE_V1=1
Also hitting this issue. Tensor parallel fix would be nice for speedy inference, but note that you can run this with --pipeline-parallel-size
I am running [unsloth/Qwen2.5-VL-72B-Instruct-bnb-4bit](https://modelscope.cn/models/unsloth/Qwen2.5-VL-72B-Instruct-bnb-4bit) succesfully, note that the dynamic quant one is not yet working, at least not for me.. You might want to [build vLLM from source or grab...