Ignace Helsen

icon location A fanatic of space sciences, automotive industry and machine learning!

Results 5 comments of Ignace Helsen

AssertionError (assert param_data.shape == loaded_weight.shape) when serving dynamic quantized models with VLLM

Same problem.

AssertionError (assert param_data.shape == loaded_weight.shape) when serving dynamic quantized models with VLLM

@Vaunorage you can still use the 16 bit to create a gguf or gptq to use in vllm

Support for Google Gemma 3

Yes please

Fix visible deprecation warning, pkg_resources deprecation

Would love for this to be merged!

support gemma 3

+1