-
We are trying to train this model for multiple video using PPO with different masks(variable number of bit rates masked).
For example, there are currently maximum 12 bitrates and some of them are ran…
-
hi, could you tell me in the discrete case , how to explore the action space, just use gumbel_softmax_sample? I find the self.exploration is not used?
-
Hello, thanks for your great work!
I want to ask you a question: In `lib/models/cell_searchs/search_cells.py, line58`, the weighted-sum in Gumbel-softmax is calculated as follows: `weigsum = sum( we…
-
how to understand magic_model
why augmentation.classifier line 158 dev_loss.backward() can update the Generator weight
-
您的论文中提到的对于mask在正向传播与反向传播时的处理,我好像没有在代码中找到对应的部分。
-
Hi, @danijar
For discrete actions, have you considered [gumbel-softmax](https://arxiv.org/abs/1611.01144), a reparameterization trick for categorical distributions? Here's my code if you haven't a…
-
Somewhat new to numpyro, though more familiar with Jax, so apologies if this is a known issue.
Modelling the boilerplate off of the baseball and time-series forcasting examples, working on a netwo…
-
Hi!
Great repo, I am glad that you implemented the code in Pytorch!
If you use exploration=True, are you really exploring? Epsilon=0.0 always, so you would never use a random selected action, or…
-
@D-X-Y Hi, I noticed the implementation of Gumbel-softmax in GDAS:
gumbels = -torch.empty_like(xins).exponential_().log()
logits = (xins.log_softmax(dim=1) + gumbels) / self.tau
…
-
In your model_search.py code you simply use `F.softmax(...)` on alphas and betas to learn the super-cell and down-sampling connections (line 275):
```python
alphas0 = F.softmax(getattr(sel…
maaft updated
4 years ago