YashRunwal

@yjh0410 Yes, I tried using Gradient Accumulation with BatchNorm layers. I don't think we can use GN layers with Gradient Accumulation. I tried using 32, 64 groups. But the result...

@yjh0410 I can get the weights for different classes. It is not an issue. But how do we assign these weights to our classes inside the gt_tensor? Example: https://discuss.pytorch.org/t/some-problems-with-weightedrandomsampler/23242/40?u=duddal

@yjh0410 How would you suggest addressing the class imbalance problem? I can't understand how do we use WeightedRandomSampler with our targets.

@yjh0410 So I changed the gt_tensor script as follows: ``` gt_tensor = np.zeros([hs, ws, num_classes + 4 + 1 + 4 + 1]) gt = [] for gt_label in range(len(label_lists)):...

@yjh0410 Sorry for tagging you here again. But what do you think of the aforementioned addition to the gt_tensor. I trained the model and the depth value should be ranging...

Wow, that was a quick reply @yjh0410 and @developer0hye :) Umm, No, I don't want to convert my Grayscale image to RGB. I want to use the 1 channel images...

@yjh0410 No problem. I will try it out and post the results here. So please don't close this thread, might be helpful for someone else. Also, 1. I have changed...

@zylo117 Thanks for responding. How many TFlops do you think an Embedded device should have in order to run Effdet in Real-time? Or, Can we tweak the architecture in order...

Thanks for the reply. Ahh damn. I want to infer real-time on Jetson. Can you perhaps redirect me to a paper or a repo that comes close to this? Or...

@gavincoder2020 Does it? I want to load 16 bit images too.