Unable to reproduce model performance using pretrained models

## ❓ Questions and Help
Hi everyone,

I have been trying to reproduce the model performance stated on the research paper but to no avail. Hence, I would like to clarify whether my command is correct.

For starters, I have downloaded the dataset from https://hatefulmemeschallenge.com/. Subsequently, I ran the following code (for Visual BERT COCO) based my understanding of the README.md:
```python
mmf_run dataset=hateful_memes \
model=visual_bert \
config=projects/hateful_memes/configs/visual_bert/from_coco.yaml \
checkpoint.resume_file=visual_bert.finetuned.hateful_memes.from_coco \
checkpoint.resume_pretrained=False \
run_type=val
```

Using this, I achieved an underwhelming result of 0.5925 for accuracy and 0.4806 ROC AUC score. This seems to be far below the performance reported in the paper...

P.S. I have also seen [Issue 926](https://github.com/facebookresearch/mmf/issues/926). Does this means the trained model checkpoint has not been updated and we should train the model from scratch?

Start with README.md and the provided mmf_run command, then inspect projects/hateful_memes/configs/visual_bert/from_coco.yaml and the pretrained checkpoint reference. Compare the reported validation metrics with Issue 926; done means confirming whether the checkpoint and command reproduce the paper’s performance or documenting the discrepancy.

Written by the indexing model from the issue text.