You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
2024-02-15 - (News) - AutoGPTQ 0.7.0 is released, with [Marlin](https://github.com/IST-DASLab/marlin) int4*fp16 matrix multiplication kernel support, with the argument use_marlin=True when loading models.
https://github.com/huggingface/optimum/blob/main/optimum/gptq/quantizer.py need a kernel choice config
Feature request
refer to https://github.com/AutoGPTQ/AutoGPTQ/blob/main/README.md
https://github.com/huggingface/optimum/blob/main/optimum/gptq/quantizer.py
need a kernel choice configMotivation
See benchmark with different autogptq kernel:
https://github.com/huggingface/optimum/blob/main/tests/benchmark/README.md
Your contribution
PR if need
The text was updated successfully, but these errors were encountered: