-
### Is this a new feature, an improvement, or a change to existing functionality?
Improvement
### How would you describe the priority of this feature request
Low (would be nice)
### Please provide…
-
We need a separate product quantization API that is decoupled from IVF but can still be composed into IVF.
Ideally this API would follow FAISS or Scikit-learn'a transformer estimators.
-
### Is this a new feature, an improvement, or a change to existing functionality?
Improvement
### How would you describe the priority of this feature request
Medium
### Please provide a clear desc…
-
File "Test.py", line 173, in
outputs, feas, _, _, _, fea_loss = model.forward(imgs[:,:3*4], update_weights, False)
ValueError: not enough values to unpack (expected 6, got 2)
how can I solv…
-
Ligature substitution (e.g. `sub A gravecomb by Agrave`) and other glyph substitution (e.g. replacing `i` with `dotlessi` when an a combining mark above follows) can currently only be defined in `feat…
-
Noticing this in a diff between feaLib and fea-rs. Format 1 subtables are much more compact, since they only store a single delta value that is applied to all items, whereas format 2 has to store per-…
-
> There's definitely something funny going on here; our GSUB is about 190kb, versus 67.5kb for fontmake.
>
> It might be that `fontc` just isn't doing 'dead lookup' analysis at all? Looking at the FE…
-
There doesn't seem to be a default constructor for nvcomp::CompressionConfig. @VasuAgrawal to add more detail if needed.
-
### Is this a duplicate?
- [x] I confirmed there appear to be no [duplicate issues](https://github.com/NVIDIA/cccl/issues) for this request and that I agree to the [Code of Conduct](CODE_OF_CONDUCT.m…
-
**Description**
@rjbs suggested on #p5p:
```
snowdrop:~$ perl -MO=Concise -E 'if ($x) { say 1 }'
8 leave[1 ref] vKP/REFC ->(end)
1 enter v ->2
2 nextstate(main 2 -e:1) v:%,us,{,fea…