[ Misc ] non-uniform quantization via `compressed-tensors` for `Llama`#6515

Merged

robertgshaw2-redhat merged 28 commits intovllm-project:mainfrom

neuralmagic:non-uniform

Jul 19, 2024

Commits on Jul 17, 2024

stash
rshaw@neuralmagic.com
committed
re-enable nonuniform for llama
rshaw@neuralmagic.com
committed
stash
rshaw@neuralmagic.com
committed
working e2e
rshaw@neuralmagic.com
committed
cleanup code a bit
rshaw@neuralmagic.com
committed
nonuniform
rshaw@neuralmagic.com
committed
fix nit
rshaw@neuralmagic.com
committed
Merge branch 'main' into non-uniform
rshaw@neuralmagic.com
committed
format
rshaw@neuralmagic.com
committed
tweak gpt2
rshaw@neuralmagic.com
committed
added test
rshaw@neuralmagic.com
committed

Commits on Jul 18, 2024

update to parent
rshaw@neuralmagic.com
committed
format
rshaw@neuralmagic.com
committed
format
rshaw@neuralmagic.com
committed
format
rshaw@neuralmagic.com
committed
format
rshaw@neuralmagic.com
committed
revert marlin change on wrong branch
rshaw@neuralmagic.com
committed
spurious chnage
rshaw@neuralmagic.com
committed
finally done
rshaw@neuralmagic.com
committed
replicated linear num lines
rshaw@neuralmagic.com
committed
remove unnessary changes
rshaw@neuralmagic.com
committed
stash
rshaw@neuralmagic.com
committed
Merge branch 'main' into non-uniform
rshaw@neuralmagic.com
committed
staged
rshaw@neuralmagic.com
committed
format
rshaw@neuralmagic.com
committed
move root back
rshaw@neuralmagic.com
committed
fix mixtral
rshaw@neuralmagic.com
committed
formatted
rshaw@neuralmagic.com
committed