Felipe

Results 7 comments of Felipe

Recommendation of problems based on the problems(/categories/tags) that user has not solved but yet popular!

How much data does stopstalk have to train a recommender system? I just found stop stalk and it seems extremely interesting!

Recommendation of problems based on the problems(/categories/tags) that user has not solved but yet popular!

Awesome. What information is there for each problem and for each user?, in order to model the features. Hoy many tagged problems are there?

Scann vs faiss

I agree is a bit cumbersome to use, but it should have (rather efficient) quantization, no? See last section of https://medium.com/@kumon/similarity-search-scann-and-4-bit-pq-ab98766b32bd

Adversarially training

Is the code using reinforce for adversarial training publicly available?

[scann]compiling error: proto_lang_toolchain rule @com_google_protobuf//:cc_toolchain: '@com_google_protobuf//:cc_toolchain' does not have mandatory provider 'ProtoInfo'

any other idea? I'm trying to compile a project with c++20, so I need a more recent bazel version.

Reverse assembly version search

Yep, that was my first thought. Thanks Ioic I'll take a look to BaGet.

Improve import speed with lazy initialization

I've also noted that when importing certain submodules of the pytorch package it slows down so much that vscodes's debugger hangs :( . An example of this is `from transformer_engine.pytorch.float8_tensor...