Tyler Troy

I have converted a custom YOLO4 model to the tensorflow saved_model format. I can successfully use this for single image inference. Now when I feed in a batch of images...

Custom Operations support

1

I am attempting to convert a tensorflow CoATnet model but the gelu activation function is not supported. I notice in the previous version of the toolkit (rknn-toolkit) there is [documentation](https://github.com/rockchip-linux/rknn-toolkit/blob/master/doc/Rockchip_Developer_Guide_RKNN_Toolkit_Custom_OP_V1.7.1_EN.pdf)...

I ran the command as specified in the docs to train on an RTX3090 using the potholes dataset as follows. ```bash python train.py --use-pretrain True --model-type p5 --dataset-type voc --dataset...

Understanding the use of taskset

1

line 139 of whisper.py includes the assertion that `assert os.sched_getaffinity(os.getpid()) == set([4, 5, 6, 7])`. What is the motivation in limiting operations to select CPUs?

Support custom_string argument in rknn export

3

### Search before asking - [x] I have searched the Ultralytics [issues](https://github.com/ultralytics/ultralytics/issues) and found no similar feature requests. ### Description When exporting an rknn model it is convenient to be...