-
I am trying to modify the current code to make it run on multi-gpu based on the tensorflow [cifar10](https://github.com/tensorflow/models/blob/master/tutorials/image/cifar10/cifar10_multi_gpu_train.py…
-
### Describe the new feature or enhancement
After a frustrating meeting with the professor and a PhD student where we wanted to take a look a the evoked responses from our pilot data, here's a requ…
-
Take args.iter_size==2 for example, I think the clipped and accumulated grads of your codes are clip(clip(grads1)+grads2), not clip(grads1+grads2), which makes more sense for me.
I haven't run the…
0HaNC updated
4 years ago
-
Hey!
First, thanks to you for your example code.
In my practice, I find there is a error in 49 line of `FCN.py` .
Personally, the `grads2 = grads[len(lr2):])` should be `grads2 = grads[len(lr1):]`. …
-
https://probot.github.io/apps/first-timers/
- We are running monthly sessions for first timers in Q1 and Q2
- This app might make us friendlier
- Would love to get grads on it
@robota-x , do you hav…
-
Optax has a function [`multi_transform`](https://optax.readthedocs.io/en/latest/api/combining_optimizers.html#optax.multi_transform), which is nice for using multiple optimizers. [See here](https://co…
-
While I run the regression test in the SW. I get an error of Protocol mismatch like following.
Can you tell me what's the possible reason and how to solve it? Thank you very much.
Running testpla…
-
hi when i compare the condition of split densification and clone densification, i found the definition of "too large gradient" is slightly between the two.
For split, the mask is generated by
htt…
-
Hello,
I am trying to build keystone and am building this version of the tool - **a1842a1ec959c4e40ffd55091795bb577894c545** - on an x86 machine, so that I can run it with gem5. I'm building on the *…
-
### 🚀 The feature, motivation and pitch
`functorch` is deprecated in favor of `torch.func` in PyTorch 2.0. The API [`functorch.make_functional`](https://pytorch.org/functorch/1.13/generated/functor…