Abeer Agrawal
Abeer Agrawal
Hi there, Currently URLLib3 is pinned to be < 2.0 - this is pretty outdated and causing issues. pyproject.toml links this as a similar bug: https://github.com/apache/arrow/commit/4295e3dd06f9f507f75ccc780eaca16a6b621a02 which doesn't exist in...
**Describe the bug** **To Reproduce** ``` export VLLM_USE_V1=0 export NCCL_DEBUG=INFO CUDA_VISIBLE_DEVICES=0 nohup trl vllm-serve --model google/gemma-3-4b-it --max-model-len 6656 --gpu-memory-utilization=0.9 --enable_prefix_caching True > outputs/vllm.txt & ``` CUDA_VISIBLE_DEVICES="1,2,3,4,5,6,7" nohup accelerate launch --config_file...