Pls provide a guide, how to use it.

#1
by S013k - opened

I've run uv pip install git+https://github.com/FayeSpica/vllm.git@glm_flash_fix on a fresh env and got errors with fa:

subprocess.CalledProcessError: Command '['cmake', '--build', '.', '-j=10', '--target=_moe_C', '--target=cumem_allocator', '--target=triton_kernels', '--target=_vllm_fa2_C', '--target=_vllm_fa3_C', '--target=_flashmla_C',
'--target=_flashmla_extension_C', '--target=_C']' returned non-zero exit status 137.

Please, provide a full installation guide/commands for lamers with tesla v100.

The error exit status 137 is a standard Linux code indicating that your process was killed by the OOM (Out of Memory) Killer.

Try the following command. This tells the build system to only use 1 concurrent jobs, significantly reducing RAM usage:

MAX_JOBS=1 uv pip install git+https://github.com/FayeSpica/vllm.git@glm_flash_fix

Sign up or log in to comment