<!---
Copyright 2020 The HuggingFace Team. All rights reserved.

Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at

    http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
-->

## SQuAD

Based on the script [`run_qa.py`](https://github.com/huggingface/transformers/blob/master/examples/question-answering/run_qa.py).

**Note:** This script only works with models that have a fast tokenizer (backed by the 🤗 Tokenizers library) as it
uses special features of those tokenizers. You can check if your favorite model has a fast tokenizer in
[this table](https://huggingface.co/transformers/index.html#bigtable), if it doesn't you can still use the old version
of the script.

The old version of this script can be found [here](https://github.com/huggingface/transformers/tree/master/examples/legacy/question-answering).
#### Fine-tuning BERT on SQuAD1.0

This example code fine-tunes BERT on the SQuAD1.0 dataset. It runs in 24 min (with BERT-base) or 68 min (with BERT-large)
on a single tesla V100 16GB.

```bash
python run_qa.py \
  --model_name_or_path bert-base-uncased \
  --dataset_name squad \
  --do_train \
  --do_eval \
  --per_device_train_batch_size 12 \
  --learning_rate 3e-5 \
  --num_train_epochs 2 \
  --max_seq_length 384 \
  --doc_stride 128 \
  --output_dir /tmp/debug_squad/
```

Training with the previously defined hyper-parameters yields the following results:

```bash
f1 = 88.52
exact_match = 81.22
```

#### Distributed training


Here is an example using distributed training on 8 V100 GPUs and Bert Whole Word Masking uncased model to reach a F1 > 93 on SQuAD1.1:

```bash
python -m torch.distributed.launch --nproc_per_node=8 ./examples/question-answering/run_squad.py \
    --model_name_or_path bert-large-uncased-whole-word-masking \
    --dataset_name squad \
    --do_train \
    --do_eval \
    --learning_rate 3e-5 \
    --num_train_epochs 2 \
    --max_seq_length 384 \
    --doc_stride 128 \
    --output_dir ./examples/models/wwm_uncased_finetuned_squad/ \
    --per_device_eval_batch_size=3   \
    --per_device_train_batch_size=3   \
```

Training with the previously defined hyper-parameters yields the following results:

```bash
f1 = 93.15
exact_match = 86.91
```

This fine-tuned model is available as a checkpoint under the reference
[`bert-large-uncased-whole-word-masking-finetuned-squad`](https://huggingface.co/bert-large-uncased-whole-word-masking-finetuned-squad).

#### Fine-tuning XLNet with beam search on SQuAD

This example code fine-tunes XLNet on both SQuAD1.0 and SQuAD2.0 dataset.

##### Command for SQuAD1.0:

```bash
python run_qa_beam_search.py \
    --model_name_or_path xlnet-large-cased \
    --dataset_name squad \
    --do_train \
    --do_eval \
    --learning_rate 3e-5 \
    --num_train_epochs 2 \
    --max_seq_length 384 \
    --doc_stride 128 \
    --output_dir ./wwm_cased_finetuned_squad/ \
    --per_device_eval_batch_size=4  \
    --per_device_train_batch_size=4   \
    --save_steps 5000
```

##### Command for SQuAD2.0:

```bash
export SQUAD_DIR=/path/to/SQUAD

python run_qa_beam_search.py \
    --model_name_or_path xlnet-large-cased \
    --dataset_name squad_v2 \
    --do_train \
    --do_eval \
    --version_2_with_negative \
    --learning_rate 3e-5 \
    --num_train_epochs 4 \
    --max_seq_length 384 \
    --doc_stride 128 \
    --output_dir ./wwm_cased_finetuned_squad/ \
    --per_device_eval_batch_size=2  \
    --per_device_train_batch_size=2   \
    --save_steps 5000
```

Larger batch size may improve the performance while costing more memory.

##### Results for SQuAD1.0 with the previously defined hyper-parameters:

```python
{
"exact": 85.45884578997162,
"f1": 92.5974600601065,
"total": 10570,
"HasAns_exact": 85.45884578997162,
"HasAns_f1": 92.59746006010651,
"HasAns_total": 10570
}
```

##### Results for SQuAD2.0 with the previously defined hyper-parameters:

```python
{
"exact": 80.4177545691906,
"f1": 84.07154997729623,
"total": 11873,
"HasAns_exact": 76.73751686909581,
"HasAns_f1": 84.05558584352873,
"HasAns_total": 5928,
"NoAns_exact": 84.0874684608915,
"NoAns_f1": 84.0874684608915,
"NoAns_total": 5945
}
```

#### Fine-tuning BERT on SQuAD1.0 with relative position embeddings

The following examples show how to fine-tune BERT models with different relative position embeddings. The BERT model 
`bert-base-uncased` was pretrained with default absolute position embeddings. We provide the following pretrained 
models which were pre-trained on the same training data (BooksCorpus and English Wikipedia) as in the BERT model 
training, but with different relative position embeddings. 

* `zhiheng-huang/bert-base-uncased-embedding-relative-key`, trained from scratch with relative embedding proposed by 
Shaw et al., [Self-Attention with Relative Position Representations](https://arxiv.org/abs/1803.02155)
* `zhiheng-huang/bert-base-uncased-embedding-relative-key-query`, trained from scratch with relative embedding method 4 
in Huang et al. [Improve Transformer Models with Better Relative Position Embeddings](https://arxiv.org/abs/2009.13658)
* `zhiheng-huang/bert-large-uncased-whole-word-masking-embedding-relative-key-query`, fine-tuned from model 
`bert-large-uncased-whole-word-masking` with 3 additional epochs with relative embedding method 4 in Huang et al. 
[Improve Transformer Models with Better Relative Position Embeddings](https://arxiv.org/abs/2009.13658)


##### Base models fine-tuning

```bash
export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
python -m torch.distributed.launch --nproc_per_node=8 ./examples/question-answering/run_squad.py \
    --model_name_or_path zhiheng-huang/bert-base-uncased-embedding-relative-key-query \
    --dataset_name squad \
    --do_train \
    --do_eval \
    --learning_rate 3e-5 \
    --num_train_epochs 2 \
    --max_seq_length 512 \
    --doc_stride 128 \
    --output_dir relative_squad \
    --per_device_eval_batch_size=60 \
    --per_device_train_batch_size=6
```
Training with the above command leads to the following results. It boosts the BERT default from f1 score of 88.52 to 90.54.

```bash
'exact': 83.6802270577105, 'f1': 90.54772098174814
```

The change of `max_seq_length` from 512 to 384 in the above command leads to the f1 score of 90.34. Replacing the above 
model `zhiheng-huang/bert-base-uncased-embedding-relative-key-query` with 
`zhiheng-huang/bert-base-uncased-embedding-relative-key` leads to the f1 score of 89.51. The changing of 8 gpus to one 
gpu training leads to the f1 score of 90.71.

##### Large models fine-tuning

```bash
export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
python -m torch.distributed.launch --nproc_per_node=8 ./examples/question-answering/run_squad.py \
    --model_name_or_path zhiheng-huang/bert-large-uncased-whole-word-masking-embedding-relative-key-query \
    --dataset_name squad \
    --do_train \
    --do_eval \
    --learning_rate 3e-5 \
    --num_train_epochs 2 \
    --max_seq_length 512 \
    --doc_stride 128 \
    --output_dir relative_squad \
    --per_gpu_eval_batch_size=6 \
    --per_gpu_train_batch_size=2 \
    --gradient_accumulation_steps 3
```
Training with the above command leads to the f1 score of 93.52, which is slightly better than the f1 score of 93.15 for 
`bert-large-uncased-whole-word-masking`.

## SQuAD with the Tensorflow Trainer

```bash
python run_tf_squad.py \
    --model_name_or_path bert-base-uncased \
    --output_dir model \
    --max_seq_length 384 \
    --num_train_epochs 2 \
    --per_gpu_train_batch_size 8 \
    --per_gpu_eval_batch_size 16 \
    --do_train \
    --logging_dir logs \    
    --logging_steps 10 \
    --learning_rate 3e-5 \
    --doc_stride 128    
```

For the moment evaluation is not available in the Tensorflow Trainer only the training.