Reorganize examples (#9010)

* Reorganize example folder

* Continue reorganization

* Change requirements for tests

* Final cleanup

* Finish regroup with tests all passing

* Copyright

* Requirements and readme

* Make a full link for the documentation

* Address review comments

* Apply suggestions from code review
Co-authored-by: Lysandre Debut <lysandre@huggingface.co>

* Add symlink

* Reorg again

* Apply suggestions from code review
Co-authored-by: Thomas Wolf <thomwolf@users.noreply.github.com>

* Adapt title

* Update to new strucutre

* Remove test

* Update READMEs
Co-authored-by: Lysandre Debut <lysandre@huggingface.co>
Co-authored-by: Thomas Wolf <thomwolf@users.noreply.github.com>

examples/longform-qa/README.md → examples/research_projects/longform-qa/README.md

 # Long Form Question Answering
 
+Author: @yjernite
+
 This folder contains the code for the Long Form Question answering [demo](http://35.226.96.115:8080/) as well as methods to train and use a fully end-to-end Long Form Question Answering system using the [🤗transformers](https://github.com/huggingface/transformers) and [🤗datasets](https://github.com/huggingface/datasets) libraries.
 
 You can use these methods to train your own system by following along the associate [notebook](https://github.com/huggingface/notebooks/blob/master/longform-qa/Long_Form_Question_Answering_with_ELI5_and_Wikipedia.ipynb) or [blog post](https://yjernite.github.io/lfqa.html).

examples/research_projects/longform-qa/requirements.txt

+datasets >= 1.1.3
+faiss-cpu
+streamlit
+elasticsearch

examples/movement-pruning/README.md → examples/research_projects/movement-pruning/README.md

 # Movement Pruning: Adaptive Sparsity by Fine-Tuning
 
+Author: @VictorSanh
+
 *Magnitude pruning is a widely used strategy for reducing model size in pure supervised learning; however, it is less effective in the transfer learning regime that has become standard for state-of-the-art natural language processing applications. We propose the use of *movement pruning*, a simple, deterministic first-order weight pruning method that is more adaptive to pretrained model fine-tuning. Experiments show that when pruning large pretrained language models, movement pruning shows significant improvements in high-sparsity regimes. When combined with distillation, the approach achieves minimal accuracy loss with down to only 3% of the model parameters:*
 
 | Fine-pruning+Distillation<br>(Teacher=BERT-base fine-tuned) | BERT base<br>fine-tuned | Remaining<br>Weights (%) | Magnitude Pruning      | L0 Regularization      | Movement Pruning       | Soft Movement Pruning          |