-
Makes it much harder to follow large IL logic in linear view
-
Refacto all the core repositories :
* Use async SQLmodel/SQLAlchemy session
-
Imagine a branching logic with True and False branch laid out Top-Bottom structure. What I wish to do is always have `True` branch on the left side and `False` on the right.
Currently, I have a funct…
-
Today, we advertise that you can use torch.export on a model that has a tensor subclasses as model params/buffers as long as you write a custom parametrization that "desugars" your subclass into the s…
-
# **Context of this change**
The preset royalty policies we will have will only cover higher priority use cases. However, different apps may want their very unique logic that can only be made from …
-
# Main todos:
- [ ] Check whether all move sorting table are correctly allocated, cleared or scaled, prepare some tests for them. Maybe clearing is not necessary
- [ ] Prepare and test better coef …
-
We need to check if the dataset is properly exponentiating as the observation instances within the dataset grows. This logic should generally standardize the order magnitude that is acceptable, for g…
-
Hi, there are some linear activation layers in my model. Since the keras-surgeon won't take linear activation layer into account, the pruning for my model failed.
Here is the code in Keras-Surgeon:
…
-
The `customBackoff` function that can be passed to `retryDelayOptions` doesn't appear to change the retry delay at all. The delay still appears to be 300ms no matter what. The message printed on scree…
-
There has been some community appetite for classification tasks #1249 #1124. Incidentally, due to the use of classification models for RLHF, we already have some of the necessary components to support…