Daniel Smilkov
Daniel Smilkov
Hi, Support for `KerasLayer` will probably not happen in the next 3 months due to the complexity of the solution. For a workaround, can you try serializing the model to...
@tafsiri before you move this PR, can you reply to my comments here. Once we wrap up the discussions, I can LGTM here, and then you can simply open a...
Also let's do the PRs one by one - this will be a learning opportunity for the code style, how ops work in TF.js etc. And it will impact how...
Yeah, that's quite a difference in performance! Let's do the `qr` first. Thanks!!
Thank you for this proposal! We just released an alpha version of the TensorFlow.js WASM backend, which was one of our most requested features. Benchmarks show 30-50% speedup with QFMA...
it should also speedup opening new datasets
This confused me too, but then I learned that github keys the cache by the branch name of your PR. When you open a PR for the first time, it...
We just added `dataset.cluster(skip_noisy_assignment=...)` (UI support too) in https://github.com/lilacai/lilac/pull/1194 When set to True, it will skip assigning noisy points to the nearest cluster to speedup clustering. This will be available...
It's `136,109` QA pairs. Took 52mins :( ```Task completed: "[local/glaive-coder] Compute signal "cluster_hdbscan" on "answer"" in 52m52s.```
Hi Mike! Thanks for filing this! Can you paste the entire URL when you are in this state. That will helps us see the exact combination of filters