Zaccharie Ramzi

I personally have been in the last few days: - I have setup W&B to run offline and be synced from a front node: see [this script](https://github.com/zaccharieramzi/submission-scripts/blob/master/jean_zay/syncall_wandb.sh). - I just...

Yes probably, and how can I change that?

Hi! Is there a workaround for the time being? Like passing `data_dir` or something like that? I would like to use [this diffuser example](https://github.com/huggingface/diffusers/tree/main/examples/unconditional_image_generation) on my cluster whose nodes are...

Apparently the failed CI test `keras-cv-accelerator-testing (keras-team-test)` is due to a lack of authorization. I am not sure I can act on it

@LukeWood I am not sure where the weights are supposed to be. [This paragraph](https://github.com/keras-team/keras-cv#pretrained-weights) points to [this file](https://github.com/keras-team/keras-cv/blob/master/examples/training/classification/imagenet/training_history.json), but there is no information regarding weights.

But the original paper doesn't specify that the convolution kernels should be intialized with "orthogonal" right?

No I think you are right, the batch normalization used in the if clause governed by `use_bnorm` is [commented out](https://github.com/cszn/DnCNN/blob/master/TrainingCodes/dncnn_keras/main_train.py#L66-L69). The batch normalization is therefore always used, no matter the...

I think one should really try to avoid modifying PyTorch's source code. A simpler solution is that provided by @GatzZ .

@richinex could be related to [this](https://github.com/google/jaxopt/discussions/322), and therefore with the fixes in https://github.com/google/jaxopt/pull/323 and https://github.com/google/jaxopt/pull/350 it might make the results consistent with that of core JAX. Maybe you can give...

@nict-wisdom do you have a snippet showing how you used the `ProfilerHook`, I am a bit struggling with it atm.