missing finetuned text features in MSRVTT (teachtext)

albanie / collaborative-experts

Video embeddings for retrieval with natural language queries

https://www.robots.ox.ac.uk/~vgg/research/collaborative-experts/

Apache License 2.0

332 stars 55 forks source link

missing finetuned text features in MSRVTT (teachtext) #29

Closed hooraid closed 2 years ago

hooraid commented 3 years ago

I successfully reproduce your code (teachtext) but cannot on MSRVTT dataset, because"gpt2-xl-finetune.pkl" is missing. I hope to fix provided MSRVTT features. Also, "raw-captions.pkl" is missing too. (but I copy same file from old[CE] data) Thanks

Tortoise17 commented 3 years ago

@hooraid can you share the script? or have you tried on other dataset or random file / data?

ioanacroi commented 3 years ago

@hooraid Sorry for the inconvenience. We have updated the MSRVTT archive. Cheers!

hooraid commented 3 years ago

@hooraid can you share the script? or have you tried on other dataset or random file / data?

I just tried scripts from the repository. I did not try a custom dataset.