cnavarrete

Results 10 comments of cnavarrete

Guys, dont use this version of pytorch YOLOv4, its a total waste of time.

Do you recommend any other yolov4 repo compatible with pytorch?

FFmpeg hangs while attempting to transcode a h264 video

Ok I had the same problem and this is how I fixed it: You need to checkout to [7baa3d8](https://github.com/jocover/jetson-ffmpeg/tree/7baa3d8232478b82799a90c37b86a986e4df1df2), also you need to apply the patch from this version. ```...

Async grab for 2 cameras results in incomplete frames being grabbed

Thanks, while doing everything you suggested surely helped the rate at which I can capture complete images, in the end I'm still recivieng some incomplete frames. I also tested 1080p@60fps...

Can't make image capture with external trigger work

Thanks that was indeed the problem and now I can grab images using an external trigger. I have a couple of questions, first how do I "turn off" the trigger...

Can't make image capture with external trigger work

Thank I will try to contact support, also what @BernardoLuck says it's pretty interesting. Looks like there's a problem when submiting contact form "connection - Error connecting to Salesforce". I...

Train.py - _pickle.UnpicklingError: STACK_GLOBAL requires str

Delete the labels.cache in each train, val and test folders. I had this exact same problem a few days ago, I had first trained a yolov5 model and that created...

TypeError: Could not build a TypeSpec for KerasTensor(type_spec=TensorSpec(shape=(None, None, 4), dtype=tf.float32, name=None), name='tf.math.truediv/truediv:0', description="created by layer 'tf.math.truediv'") of unsupported type <class 'keras.engine.keras_tensor.KerasTensor'>.

Which version of keras and tensorflow are you using? I recommend using conda for this knd of stuff since it's likely you are gonna have a hard time porting this...

Train.py didn't start training

That's weird. Can you test it outside of the conda env?

Target layer for Pytorch's official implementation of Swin Transformer v2?

> > The grad cam has been very useful for most models, but I'm at loss on how to use it for the pytorch implementation of the Swin Transformer v2....

🐛 IllegalStateException: maxImages (6) has already been acquired, call #close before acquiring more.

@TryRevilo Could you ever solve this?