Amanpreet Singh

icon indicating a website link https://apsdehal.in

icon company CTO @ContextualAI. Past @huggingface @facebookresearch icon location San Francisco icon location Building @ContextualAI. Working on intersection of NLP, Vision and RL. Past @facebookresearch, @huggingface

Results 49 comments of Amanpreet Singh

With dataloader RSS memory consumed by HF datasets monotonically increases

It is hard to say if it is directly reproducible in this setup. Maybe it is specific to the images stored in the CM4 case which cause a memory leak....

Tests

That are not tests -_-

Temperature scaled sampling for multi dataset

@parthduggal Thanks for your interest. This can be added as a form of iteration strategy. Take a look at https://github.com/facebookresearch/mmf/blob/master/mmf/datasets/iteration_strategies.py

mmf hateful memes challenge dataset

Hi, The data is still available here: https://hatefulmemeschallenge.com/#download You might need to make minor changes to extraction script to get it working with recent format.

features from x-152pp pretrained model doesn't recreate the "movie_mcan.grid.vqa2_vg" test-dev results

Hi, Can you post your full logs here?

features from x-152pp pretrained model doesn't recreate the "movie_mcan.grid.vqa2_vg" test-dev results

When you ran the `mmf_predict` command that you mentioned, some logs must have been produced to stdout. Can you provide those logs?

features from x-152pp pretrained model doesn't recreate the "movie_mcan.grid.vqa2_vg" test-dev results

Running it now, will let you know the results.

features from x-152pp pretrained model doesn't recreate the "movie_mcan.grid.vqa2_vg" test-dev results

Hi, here's what I got: ``` [{"test-dev": {"yes/no": 89.13, "number": 58.33, "other": 64.5, "overall": 73.94}}] ``` I would open up an issue on grids-feat-vqa repo to understand if you are...

features from x-152pp pretrained model doesn't recreate the "movie_mcan.grid.vqa2_vg" test-dev results

For basic sanity check, you should first try to match the features you are extracting with the ones that are downloaded by MMF. Also, if there are minor difference in...

errors while running pytests .\test

@LeungWaiHo can you try install MMF properly following the instructions in the docs? It looks like you are missing the transformers package which should have been installed with MMF dependencies.