Mithrillion

Why differently shaped tensors can't be concatenated with rearrange?

4

Currently, concatenation as in the example is done by calling `stack_on_zeroth_dimension()` first then rearranging the tensor into the appropriate shape. However, most `backend.stack()` requires that all except the stacked dimension...

Is it possible to expose the exp map to the Python API via a logsignature_to_signature function?

2

Somehow, I found myself needing to map a logsignature back to the full signature (where the logsignature can be in either the tensor algebra or a compressed form). It appears...

Bypassing ApplyInstantID Node Can Ruin Conditioning

8

Issue screenshot: Using a slightly modified default workflow, we can show that bypassing ApplyInstantID node does not maintain the original behaviour but rather degrades the quality of conditioning.  Expected...

InstantID masking is not properly applied when used alongside differential diffusion

2

Screenshot:  Issue description: When using masking in both InstantID and masked differential diffusion, the noisy diffusion output from regions outside of masked areas appear to be pasted over the...

Result not recorded if a search term converges to a single entry

2

It appears that currently the scraper is not able to save results from search terms that only return one entry (which leads to that entry being displayed directly the UI...

It appears there are some issues with the user group setup for the image. When run with `--user root` the image is able to load, however without it the image...

### Describe the issue: Everthing works fine if using GPU for sampling. However, if I try to sample on the CPU by `os.environ["JAX_PLATFORM_NAME"] = "cpu"` and set cores to >1,...