Chanchana Sornsoontorn

@sourabpramanik That isn't ideal because I want to check remaining values often without modifying the limit. For example, I might want to display this remaining value on every page refresh,...

Let me give another example use case of why `getRemainingQuota` is important. Suppose you limit the user to 100 requests per hour based on his email address. Some users will...

I have tested the code with dummy 2 circle data and predict their positions and it works.  The red and green dots are the predictions of the model. The...

But I'm not sure whether `js_reg_loss` will work with multi-channel though because I haven't tried it yet. As it's not trivial for me to use a custom loss of this...

You can use it fine. Except there is no regularization loss yet. Because I'm not familiar with it. So don't use the regularization like `js_reg_loss`. But I'm sure the multi-channel...

@hjpulkki Dataset can be created simply by creating a black image, then draw a circle using `cv2.circle()` function at a random location on the image. Use that random location as...

It seems to me that you need to prepare the label in the range (-1, 1). Excluding -1 and 1 with the paper approach. But I think that these values...

For example, I changed these lines: https://github.com/ashwhall/dsnt/blob/2a6761a1cf0aef8fe7c875ce4a6225631072775a/dsnt.py#L26-L30 To be like this: ```py # Build the DSNT x, y matrices # dsnt_x = tf.tile([[(2 * tf.range(1, width+1) - (width + 1))...

Check my PR #8, it has an option for the user to specify 0to1 or -1to1.

@lllyasviel can you share hyperparameters you used for training e.g. batch size, effective batch size, number of GPUs, number of worker nodes, learning rate, number of training steps, etc? I...