Damien Menigaux

Results 19 issues of Damien Menigaux

It'd be very interesting, especially since 3d models are part of torchvision model zoo.

Solves https://github.com/inducer/pycuda/issues/319 I'm not exactly sure my work is proper, I'm more a python guy than C++. Can you share your thoughts @inducer please ?

I think I'm able to say that stream priority isn't implemented in pycuda right now. Although Cuda gives the possibility to allocate priorities to Streams, as pointed out here :...

**Description** I'm converting a pytorch model to torchscript with or without fp16 precision, and I get much slower triton inference when using FP16, even though a torchscript benchmark shows it...

I've been looking at this for the past few days and I can't seem to figure out how I can use this to train on any other dataset than Thumos...

Hi. I'm currently looking at your code. It doesn't produce a bug on 3d convolutions but I wondered if it would eventually work (one epoch with 3d convolution context is...

New error on my gccp instances getting data from s3 bucket. ``` user_name@instance~$ s4cmd -c32 get -r s3:/my-bucket/my-data/ [Exception] __init__() got an unexpected keyword argument 'cert_file' [Thread Failure] __init__() got...

Are there any plans to adapt this work to convolutionnal layers in tensorflow ?

Hi. Thank you for your greatly interesting work ! I know that a huge chunk of the industry uses pytorch for research, and tensorflow for deployment. So making this useful...

Thanks for your error report and we appreciate it a lot. **Describe the bug** pytorch2onnx.py is broken with available models **Reproduction** ``` python tools/pytorch2onnx.py DetectoRS_mstrain_400_1200_r50_40e.py DetectoRS_R50-0f1c8080.pth --out onnx_model.onnx ``` **Environment**...