Will Reynolds
willowill5
·
AI & ML interests
LLMs, Talking Head Animation
Organizations
None yet
willowill5's activity
OOM with vllm
#48 opened 7 months ago
by
willowill5
![](https://cdn-avatars.huggingface.co/v1/production/uploads/62acc1f1eda69b28bb64c39d/tmMMhkSyHc6HXDyhE7-jj.jpeg)
vLLM out of memory
2
#2 opened 8 months ago
by
cfrancois7
![](https://cdn-avatars.huggingface.co/v1/production/uploads/624ab19d93d46cf4a090da4b/O4c4XDpYiUSWtDOjrzkTa.jpeg)
OOM on RTX 3090 with vLLM
#1 opened 8 months ago
by
willowill5
![](https://cdn-avatars.huggingface.co/v1/production/uploads/62acc1f1eda69b28bb64c39d/tmMMhkSyHc6HXDyhE7-jj.jpeg)
Quantization not recognized, even when building VLLM from source
2
#1 opened 9 months ago
by
willowill5
![](https://cdn-avatars.huggingface.co/v1/production/uploads/62acc1f1eda69b28bb64c39d/tmMMhkSyHc6HXDyhE7-jj.jpeg)
very slow inference speed on 2x A100 80GB with 4-bit (main branch)
#6 opened 9 months ago
by
willowill5
![](https://cdn-avatars.huggingface.co/v1/production/uploads/62acc1f1eda69b28bb64c39d/tmMMhkSyHc6HXDyhE7-jj.jpeg)