Federico Bianchi

icon indicating a website link https://federicobianchi.io

icon company Together AI icon location San Francisco icon location Senior ML Scientist at TogetherAI. Prev. Post-Doc at Stanford and Stanford NLP. Also doing some open source stuff.

Results 54 comments of Federico Bianchi

Korean Language Support

Hello @StrangeFate, this looks great, I'd like to introduce this in a specific location of the source and to install this using a [selective install](https://stackoverflow.com/questions/5226311/installing-specific-package-versions-with-pip). Point is, I'd also like...

ctm.save crashed when training_dataset is somewhat large

Thanks! Let me label this as a bug. Might make sense to remove the dataset in a future version of the model

Parallel GPU training

Hi! Currently, we do not support parallel GPU training, I'm sorry. `DataParallel` shouldn't be too difficult to setup, but I have to explore better the thing since I do not...

Covering other langauges

Thanks!

Covering other langauges

Check if [this](https://github.com/MilaNLProc/contextualized-topic-models/blob/master/contextualized_topic_models/utils/preprocessing.py#L59) covers your usecase, otherwise I can try to adapt it to yours :)

vocab attribute used for CoherenceWordEmbeddings removed in gensim 4.x

Thanks! This probably goes with #113. We need to update all our dependencies. Thanks a lot!

Deprecation Warnings

Hello @TudorAndrei ! Thanks, guess we should update our dependencies

Variable naming issues

Thanks, that make sense! Do you want to open a PR for that?

Discrepancy between Hugging Face and fashion-clip

I am surprised because both methods seem to use the same transformation, but I'll take a look! thanks!!

Discrepancy between Hugging Face and fashion-clip

Hi @anilsathyan7! I am not sure how the UI computes the score; in the meantime, I have run your example on both the original HF API and our internal wrapper...