AttributeError: 'NoneType' object has no attribute 'format'

[miltosc]

Hello i get the following error when running python 2_dreaming_time.py -i frames_dir -o processed_frames_dir --model_path models/bvlc_googlenet/ --model_name bvlc_googlenet.caffemodel --gpu 0 --extract 0 -it png

In colab i get the error Check failed: error == cudaSuccess (48 vs. 0) no kernel image is available for execution on the device

Locally i get the error im2col.cu:61] Check failed: error == cudaSuccess (8 vs. 0) invalid device function

if i run it without --gpu 0 to use CPU after some modifications it works but very slow wish i could make it work with GPU

Thanks in advance for your help.

[tux2603]

What GPU do you have? If nvidia, are you using the proprietary drivers or nouveau?

[miltosc]

Hello, i use the proprietary drivers. I finally did it. Is there a model without only animals ? thanks.

[tux2603]

There are a couple of different options available in the caffe model zoo. It's also possible to train your own models, but that involves a lot of computational resources and a decently large data set of images

[miltosc]

I know and understand that about computational resources. for now i only have an 2080 nvidia. Thank you for your help i really appreciate.

[tux2603]

No problem. This project seems to no longer be supported, so I'm rebuilding it from scratch in tux2603/DeepDream