-
# vanishing and exploding gradient / sensitivity
- (**must see**) X. Glorot and Y. Bengio. Understanding the difficulty of trainingdeep feedforward neural networks. InAISTATS, 2010.
- (**must see**)…
-
I found komm.TerminatedConvolutionalCode.decode() only 1/2 decoding, cannot achieve 3/4 decoding. Excuse me, komm.TerminatedConvolutionalCode.decode can achieve similar to the function of comm.Viter…
-
I know this may looks a bit lame but seems difficult to understand how does the regularizer works in this case, as all the example I saw online has a regularizer = None.
What should i do if I want …
-
2025 changes to PathPlannerLib will include outputting torque-current feedforwards from path following commands and a setpoint generator utility. It appears that the Phoenix 6 Swerve mechanism's Swerv…
-
I might be missing something but in models like resnet the receptive field of each layer might not depend only the single layer that was computed before the current layer. It seems to me that this cod…
-
In an effort to make the standard feedforward networks more useful, we need to implement momentum and regularization (with an arbitrary regularization function -- l1, l2), which should improve learnin…
-
https://github.com/DeepBlueRobotics/RobotCode2024/blob/864bd307aef11cefdb3b42233551b75dce458c0b/src/main/java/org/carlmontrobotics/subsystems/Arm.java#L163
The kG portion of the feedforward should …
-
# Description
This is what we reproduced:
![image](https://user-images.githubusercontent.com/36265636/155481453-09c504ca-0b2f-46cd-8963-e9c2cd74aba6.png)
This is the result in the paper:
![ima…
-
### 🐛 Describe the bug
# reproduce the bug
@mstebelev found out that memory efficient attention kernel on float32 cuda tensors gives nan gradients despite inputs and incoming gradient are reaso…
-
Hi, thank you for your awesome pruning work.
I have had hard time to fix an issue.
I just followed the sample code you offers in #6.
After pruning, "pruend_macs, pruned_params = tp.utils.count_op…