-
## Describe the bug
Tensordictv0.6,0 was released in last night, but this is not compatible with rl4co.
https://github.com/pytorch/tensordict/releases/tag/v0.6.0
## To Reproduce
```bash
cd rl…
-
## Describe the bug
Building a `ProbabilisticActor` module with `return_log_prob=True` makes it write the log-probability of the distribution sample in a given tensordict. However, unlike the othe…
-
## Motivation
Ensembled modules should also support optimization with an API consistent with ordinary modules. However, passing `EnsembledModule.parameters()` to an optimizer, as usual, does not yi…
-
### Current state of data collectors
We currently have parallel data collectors, that can handle a single or multiple environments running serially or in parallel.
Say one has access to N cpus to ru…
-
## Describe the bug
When training on `PettingZoo/MultiWalker-v9` with `Multi-Agent Soft Actor-Critic`, **all** losses (`loss_actor`, `loss_qvalue`, `loss_alpha`) explode after ~1M environment steps…
-
## Motivation
### 1. Consistent style for `torch.nn.modules.loss.*Loss`
In `torch.nn.modules.loss`, there are many `*Loss` subclassing `nn.Module`. The `Loss.__init__()` does not takes other `nn…
-
## Motivation
Twin Q/ensemble Q functions are used in many RL algorithms and mitigate Q overestimation. My understanding is that TorchRL only deals with ensembles in the loss functions. This is fine …
-
Hi,
Thank you for the nice work. I have noticed that your tutorial has a sim2real project that is trained with crazyfile 2.1. Because the crazy file(crazyswarm) is developed based on the gazebo, …
-
### Current Tensorclass interface
To define a Tensorclass, we could write:
```python
@tensorclass
class MyTensorclass:
foo: Tensor
bar: Tensor
```
The `@tensorclass` decorator then gen…
-
Hi there 👋🏼
First of all, great job on your work, it has served as an inspiration to our research group!
We are actively developing [RL4CO](https://github.com/kaist-silab/rl4co), a library for all…