fix(vllm): drop flash-attn wheel to avoid torch 2.10 ABI mismatch

[richiejp]

The pinned flash-attn 2.8.3+cu12torch2.7 wheel breaks at import time
once vllm 0.19.1 upgrades torch to its hard-pinned 2.10.0:

ImportError: .../flash_attn_2_cuda...so: undefined symbol:
_ZN3c104cuda29c10_cuda_check_implementationEiPKcS2_ib

That C10 CUDA symbol is libtorch-version-specific. Dao-AILab has not yet
published flash-attn wheels for torch 2.10 -- the latest release (2.8.3)
tops out at torch 2.8 -- so any wheel pinned here is silently ABI-broken
the moment vllm completes its install.

vllm 0.19.1 lists flashinfer-python==0.6.6 as a hard dep, which already
covers the attention path. The only other use of flash-attn in vllm is
the rotary apply_rotary import in
vllm/model_executor/layers/rotary_embedding/common.py, which is guarded
by find_spec("flash_attn") and falls back cleanly when absent.

Also unpin torch in requirements-cublas12.txt: the 2.7.0 pin only
existed to give the flash-attn wheel a matching torch to link against.
With flash-attn gone, vllm's own torch==2.10.0 dep is the binding
constraint regardless of what we put here.

Notes for Reviewers

I don't know what are the performance implications, but we need them to update flash-attn practically speaking.

Signed commits

• Yes, I signed my commits.

[mudler]

mh interesting, we should check closer as not having flash attn is a huge slowdown

[richiejp]

mh interesting, we should check closer as not having flash attn is a huge slowdown

oh, it's still there, but implemented in flashinfer which seems to be the default in vLLM.