mattam301

Results 7 comments of mattam301

Implementation of inst2vec-imm is not available for Device Mapping task

Are you intend to upload the immediate version? Or can you suggest me what to do to reproduce that version? Thanks in advance

NOT_IMPLEMENTED : Could not find an implementation for ConvInteger(10) node with name 'Conv_0_quant'

> I am able to do quantization with: > > ``` > model_fp32 = 'model_Segformer.onnx' > model_quant = "model_dynamic_quant.onnx" > quantized_model = quantize_dynamic(model_fp32, model_quant, weight_type=QuantType.QUInt8) > ``` > > instead...

No module named 'awq_inference_engine'

I have the same issue, what should I do to resolve the problem?

2 bit AWQ results?

Hi, I am interested in that question too, since I have successfully reproduce int4 and int3 results in the paper, but when it came to int2, the ppl is totally...

AWQ and VILA dependency compatible issue

I encountered the same problem, the installation of VILA now need updating

"Expected all tensors to be on the same device" when running "Perform AWQ search" on Llama3

Thanks a lot!

Setup.py file - Version Issues

I supposed that he has the same issue as me when setting up AWQ kernel on the machine with CUDA version 11.2: run: python setup.py install get: RuntimeError: The detected...