Create README.md (#5217)

electra_large_discriminator_squad2_512 Question Answering LM

model_cards/ahotrod/electra_large_discriminator_squad2_512/README.md

+## ELECTRA_large_discriminator language model fine-tuned on SQuAD2.0
+
+### with the following results:
+
+```
+  "exact": 87.09677419354838,
+  "f1": 89.98343832723452,
+  "total": 11873,
+  "HasAns_exact": 84.66599190283401,
+  "HasAns_f1": 90.44759839056285,
+  "HasAns_total": 5928,
+  "NoAns_exact": 89.52060555088309,
+  "NoAns_f1": 89.52060555088309,
+  "NoAns_total": 5945,
+  "best_exact": 87.09677419354838,
+  "best_exact_thresh": 0.0,
+  "best_f1": 89.98343832723432,
+  "best_f1_thresh": 0.0
+```
+### from script:
+```
+python ${EXAMPLES}/run_squad.py \
+  --model_type electra \
+  --model_name_or_path google/electra-large-discriminator \
+  --do_train \
+  --do_eval \
+  --train_file ${SQUAD}/train-v2.0.json \
+  --predict_file ${SQUAD}/dev-v2.0.json \
+  --version_2_with_negative \
+  --do_lower_case \
+  --num_train_epochs 3 \
+  --warmup_steps 306 \
+  --weight_decay 0.01 \
+  --learning_rate 3e-5 \
+  --max_grad_norm 0.5 \
+  --adam_epsilon 1e-6 \
+  --max_seq_length 512 \
+  --doc_stride 128 \
+  --per_gpu_train_batch_size 8 \
+  --gradient_accumulation_steps 16 \
+  --per_gpu_eval_batch_size 128 \
+  --fp16 \
+  --fp16_opt_level O1 \
+  --threads 12 \
+  --logging_steps 50 \
+  --save_steps 1000 \
+  --overwrite_output_dir \
+  --output_dir ${MODEL_PATH}
+```
+### using the following system & software:
+```
+Transformers: 2.11.0
+PyTorch: 1.5.0
+TensorFlow: 2.2.0
+Python: 3.8.1
+OS/Platform: Linux-5.3.0-59-generic-x86_64-with-glibc2.10
+CPU/GPU: Intel i9-9900K / NVIDIA Titan RTX 24GB
+```