Eduardo Hirata-Miyasaki

Results 37 comments of Eduardo Hirata-Miyasaki

I merged #114. @ziw-liu feel free to make the changes

From the discussion with @ziw-liu, we will move forward by adding temporarily `UNeXt2_2D` as an alias to fcmae for the 0.2.0 release. Merging the architectures right now is not feasible...

Re-tested this. Seems to pass CI after merging #244. @ziw-liu

closing in favor of #273

I find it more use when it's decompressed rather than compressed. We can report both if needed. I think zarr.array does `nbytes_stored`. What do you guys think?

ended up adding `uncompressed size [GB]`

Using the prefetch=4 vs prefetch=2 has no effect on the training speed for the neuromast VS training. Here we are mostly limited by CPU->GPU pipes.

we will make a new folder under the `examples/configs`. We will add a text file that ask the users to use the ones in the repo.

As discussed, the HF model is pinned to use