Uploaded Manually due to github not working

README.md

@@ -23,4 +23,12 @@ I will not be using Docker for the purpose of this project so everything will be
 
 Milestone-2:
 The link below will take you to the hugging face space for the sentiment analysis web app
-https://huggingface.co/spaces/MJ3128/CS-GY-6613-Project
+https://huggingface.co/spaces/MJ3128/CS-GY-6613-Project
+
+Milestone-3:
+The same link will take you to the viability page. The sentiment analysis code has be commented out for convenience.
+https://huggingface.co/spaces/MJ3128/CS-GY-6613-Project
+
+Milestone-4:
+Landing page for the project, contains a video demo. Demo can also be found in the repository.
+https://sites.google.com/nyu.edu/mj3128-ai-project/home

app.py

@@ -11,8 +11,9 @@ if "score" not in st.session_state:
 
 
 def get_patent_score(pipeline, abstract, claims):
-    abstract_score = pipeline(abstract)
-    claims_score = pipeline(claims)
+    abstract_score = pipeline(
+        abstract, pad_to_max_length=True, truncation=True)
+    claims_score = pipeline(claims, pad_to_max_length=True, truncation=True)
     abstract_label = abstract_score[0]["label"]
     claims_label = claims_score[0]["label"]
     st.session_state.score = "{:.2f}".format(
@@ -21,7 +22,7 @@ def get_patent_score(pipeline, abstract, claims):
     if abstract_label == claims_label:
         st.session_state.viability = abstract_label
     else:
-        if abstract_score[0]["score"] > claims_score[0]["label"]:
+        if abstract_score[0]["score"] > claims_score[0]["score"]:
             st.session_state.viability = abstract_label
         else:
             st.session_state.viability = claims_label
@@ -29,7 +30,8 @@ def get_patent_score(pipeline, abstract, claims):
 
 checkpoint_file = "./checkpoint-3024"
 model = AutoModelForSequenceClassification.from_pretrained(checkpoint_file)
-tokenizer = AutoTokenizer.from_pretrained(checkpoint_file)
+tokenizer = AutoTokenizer.from_pretrained(
+    checkpoint_file, pad_to_max_length=True)
 pipeline = pipeline("text-classification", model=model, tokenizer=tokenizer)
 
 dataset_dict = load_dataset('HUPD/hupd',
@@ -70,6 +72,8 @@ st.button("Check Viability", on_click=get_patent_score,
 st.markdown(body="Outcome: {}, Score: {}%".format(
     st.session_state.viability, st.session_state.score))
 
+get_patent_score(pipeline=pipeline, abstract=abstract, claims=claims)
+
 # Milestone-2
 # if "sentiment" not in st.session_state:
 #     st.session_state.sentiment = ""