-
Hi guys!
I am currently getting 31/30/27 (mAP-50) and 55/53/49.5 (mAP) with this implementation, which is a bit lower than what the paper claims. I was wondering if anyone else experienced this and…
-
Is there any documentation for usage to build a network?
I want to try to implement some simple network based on for example MNIST dataset.
If there is no documentation, i think we can write on…
-
The method you propose for using `recompute_grad` is not working, except for the simplest case where all layers in the model are recomputed except the input and output layers. All other cases (e.g. wh…
ghost updated
5 years ago
-
Superresolution is a key part of our workflows: many images, particularly if cropped down to a face or figure, are just not big enough to train natively at a resolution like 512px or 1024px. Padding t…
gwern updated
3 years ago
-
Hello, do you have any example on the use of this layer?
Is this layer returning a 3D tensor?
I have tried the following example:
```
reticulate::use_condaenv("r-tensorflow", required = TRUE)
l…
-
To this day, there are no reliable benchmarks for "real-world models" across frameworks (Keras, PyTorch, JAX/Flax). A "real-world model" is the kind of model that is actually produced and used by data…
-
So I have been trying to build https://raw.githubusercontent.com/AlexeyAB/darknet/master/cfg/csresnext50-panet-spp-original-optimal.cfg using tensorflow-2.0
What is mosaic/cutmix in this file and h…
-
There's very much a question of whether this is worth the leap for us in the short term, but either way it's worth documenting what this might involve.
First of all, we use scikit-learn for running…
-
Re your MobileVit2, these two norms are not equivalent and it would be misleading to call it LayerNorm2d as the group norm w/ groups=1 is not equivalent. 'LayerNorm2d' is already used elsewhere in oth…
-
Hi,
I was trying to test the model but I get the following errors:
Using TensorFlow backend.
/home/cm72ein/Project/DeepMotion/deep-motion-master/src/FI_unet.py:59: UserWarning: Update your `Conv2…