-
In the paper The LambdaLoss Framework for Ranking Metric Optimization, the loss is optimized by EM algorithm. However, I don't see the EM algorithm anywhere in the tf ranking code.
Please explain. M…
-
Hi,
on the implementation/documentation of the callbacks, there is one feature which is unclear for newcomers (I'm still unsure after navigating the source). There is the flag train to know if we'r…
-
I think it would be useful to collect user experience / discuss / get some help about the high-level meaning / use of the PPO meta-parameters.
This may be a more general discussion than just tensor…
-
- [x] Write about Hyperparameter v/s parameter: https://en.wikipedia.org/wiki/Hyperparameter_(machine_learning)
- [x] Write about Thompson sampling
- [x] Motivate grid search expensive computation c…
-
According to figure 2 of the paper, the output of Embedder is passes thorugh the MLP(maybe Mapping Network in StyleGAN) and enters the generator, but in this code, there is no MLP(mapping network). D…
-
## Description
During the training process, I tried to change the learning rate to 0.0 after 200 steps, and I found that the accuracy experiences a dropoff. Why is that?
The following figure show…
-
Hi @ageron,
I'm following the examples in Chapter 9 an everything it is going fine until the Gradient Descent optimization with placeholders. At this point, when I compute the MSE the output value …
-
## 🐛 Bug
I'm trying to run a simple model, from an example from the pytorch [website](https://pytorch.org/tutorials/beginner/pytorch_with_examples.html). However I encounter a really strange bug, w…
-
FailedPreconditionError (see above for traceback): Attempting to use uninitialized value con1/bias
[[node con1/bias/read (defined at c:/users/tran thi diem/documents/diem/research/reinforcem…
-
In certain conditions calling CompuationGraph#output(false, data) causes following exception:
`java.lang.IllegalStateException: Mismatched shapes (shape = [2, 2], row vector shape =[4, 1])
at org…