lshaw8317
lshaw8317
#18 is related. I also had the same error. Adding `global seed_worker` to `Diffusion.sample_sequence` in diffusion.py fails to resolve issue: ``` Traceback (most recent call last): File "", line 1,...
My CUDA out ot memory error was solved by going into the config yaml file and reducing the batch_size, although was a slightly different error. Running 68GB GPU.
Hi, thanks for bringing this to our attention. It seems to be necessary to update numexpr (to version 2.11.0) in order to use the latest version of NumPy with blosc2....
Could you provide an example of this to reproduce behaviour?
> Fixed in [#391](https://github.com/Blosc/python-blosc2/pull/391) This is a separate issue to the chaining of lazy expressions, it relates to differing data types of operands.
For implementing fancy indexing in slice (i.e. return blosc array not numpy array), one should return compressed output of ``get_fselection`` if the result can fit (uncompressed) in memory. If not,...
See PR #467
Can maybe be implemented using expand_dims (see PR #427)
Seems like the remaining functions we need are: - [ ] ``tril`` - [ ] ``triu`` - [ ] ``flip`` - [ ] ``repeat`` - [ ] ``roll`` - [...
Relatedly, it would be good if default behaviour of reductions acting on blosc2 arrays were to return blosc2 arrays (not numpy arrays, as happens currently).