Updated Parameters

README.md

@@ -16,31 +16,34 @@ datasets:
 
 ## Hyperparameters
 ```
-batch_size = 8
-n_epochs = 6
+batch_size=8
+n_epochs=6
 base_LM_model = "deepset/roberta-base-squad2"
-max_seq_len = 386
+max_seq_len=386
 doc_stride=128
-learning_rate = 1.5e-5
-warmup_proportion = 0.2
-batch_size=8
+learning_rate=1.5e-5
+adam_epsilon=1e-5
+adam_beta1=0.95
+adam_beta2=0.999
+warmup_steps=100
+weight_decay=0.01
 n_best_size=20
 max_answer_length=30
 min_null_score=7.0
-CLS_threshold=-3
 ```
+##### There is a special threshold value CLS_threshold=-3 used to more accurately identify no answers [Logic will be available in GitHub Repo [TBD]
 
 ## Performance
 ```
 "exact": 81.192622
-"f1": 83.95408
+"f1":    83.95408
 "total": 11873
 "HasAns_exact": 74.190283
-"HasAns_f1": 79.721119
+"HasAns_f1":    79.721119
 "HasAns_total": 5928
-"NoAns_exact": 88.174937
-"NoAns_f1": 88.174937
-"NoAns_total": 5945
+"NoAns_exact":  88.174937
+"NoAns_f1":     88.174937
+"NoAns_total":  5945
 ```
 
 ## Usage
@@ -62,4 +65,7 @@ print(res)
 # b) Load model & tokenizer
 model = AutoModelForQuestionAnswering.from_pretrained(model_name)
 tokenizer = AutoTokenizer.from_pretrained(model_name)
-```
+```
+
+## Authors
+Premal Matalia