Update app.py

app.py

@@ -1,6 +1,9 @@
-# https://huggingface.co/transformers/main_classes/pipelines.html
-# https://huggingface.co/models?filter=conversational
-
+# Benchmarks: NT, Why is blood important?
+#model_name = "deepset/roberta-base-squad2"       # 180
+#model_name = "deepset/deberta-v3-large-squad2"  # est. 4X
+model_name = "deepset/tinyroberta-squad2"       # 86
+#model_name = "deepset/minilm-uncased-squad2"    # 96
+#model_name = "deepset/electra-base-squad2"      # 185 (nice wordy results)
 
 
 # Install Dependences
@@ -9,22 +12,17 @@
 # !pip install ipywidgets 
 # !pip install gradio  # see setup for installing gradio
 
-# Import Dependencies
-from transformers import pipeline 
 import gradio as gr
+from transformers import AutoModelForQuestionAnswering, AutoTokenizer, pipeline
 
-# Create the Q&A pipeline
-nlp = pipeline('question-answering', model='deepset/roberta-base-squad2', tokenizer='deepset/roberta-base-squad2')
-#nlp = pipeline('question-answering', model='bert-large-uncased-whole-word-masking-finetuned-squad ', tokenizer='bert-large-uncased-whole-word-masking-finetuned-squad ')
-#nlp = pipeline("question-answering", model='distilbert-base-cased-distilled-squad')
-#nlp = pipeline("question-answering", model='distilbert-base-uncased-distilled-squad')
+nlp = pipeline('question-answering', model=model_name, tokenizer=model_name)
 
 def question_answer(context_filename, question):
     """Produce a NLP response based on the input text filename and question."""
     with open(context_filename) as f:
         context = f.read()
-    nlp_input =  {'question': question, 'context': context}
-    result = nlp(nlp_input)
+        nlp_input =  {'question': question, 'context': context}
+        result = nlp(nlp_input)
     return result['answer']
 
 demo = gr.Interface(