examples

app.py

@@ -54,70 +54,172 @@ def infer(
     print(f"recv: {datetime.now()}")
     
     for stop_word in stop:
-        if stop_word in generated_text:
+        if stop_word != '' and stop_word in generated_text:
             generated_text = generated_text[:generated_text.find(stop_word)]
-            
-    st.session_state.updated = True
     
     return generated_text
 
 
 def set_preset():
-    if st.session_state.preset == "Classification":
+    if st.session_state.preset == "Question Answering":
         
-        if not st.session_state.updated:
-            st.session_state.prompt = '''Please classify the given sentence.
-Possible labels:
-1. <label_0>
-2. <label_1>
+        st.session_state.prompt = '''
+Please answer the following question:
 
-Input: <sentence_0>
-Label: <label_0>
+Question: What is the capital of Canada?
+Answer: Ottawa
 
-Input: <sentence_1>
-Label:'''
+Question: What is the currency of Switzerland?
+Answer: Swiss franc
+
+Question: In which country is Wisconsin located?
+Answer:
+'''.strip()
         st.session_state.temperature = "0.0"
         st.session_state.top_p = "1.0"
-        st.session_state.max_new_tokens = "10"
+        st.session_state.max_new_tokens = "5"
+        st.session_state.stop = r'\n'
         
-    elif st.session_state.preset == "Generation":
+    elif st.session_state.preset == "Sentiment Analysis":
         
-        if not st.session_state.updated:
-            st.session_state.prompt = '''Please write a story given keywords.
+        st.session_state.prompt = '''
+Label the tweets as either "positive", "negative", "mixed", or "neutral":
+
+Tweet: I can say that there isn't anything I would change.
+Label: positive
+
+Tweet: I'm not sure about this.
+Label: neutral
 
-Input: bear, honey
-Story: Once upon a time,'''
+Tweet: I liked some parts but I didn't like other parts.
+Label: mixed
+
+Tweet: I think the background image could have been better.
+Label: negative
+
+Tweet I really like it.
+Label:
+'''.strip()
         st.session_state.temperature = "0.0"
-        st.session_state.top_p = "0.9"
-        st.session_state.max_new_tokens = "100"
-    
-    else:
-        pass
-    
-    
-def main():
+        st.session_state.top_p = "1.0"
+        st.session_state.max_new_tokens = "2"
+        st.session_state.stop = r'\n'
+        
+    elif st.session_state.preset == "Topic Classification":
+        
+        st.session_state.prompt = '''
+Given a news article, classify its topic.
+Possible labels: 1. World 2. Sports 3. Business 4. Sci/Tech
 
-    if 'preset' not in st.session_state:
-        st.session_state.preset = "Classification"
+Article: A nearby star thought to harbor comets and asteroids now appears to be home to planets, too.
+Label: Sci/Tech
 
-    if 'prompt' not in st.session_state:
-        st.session_state.prompt = "Please answer the following question:\n\nQuestion: In which country is Zurich located?\nAnswer:"
+Article: Soaring crude prices plus worries about the economy and the outlook for earnings are expected to hang over the stock market next week during the depth of the summer doldrums.
+Label: Business
 
-    if 'temperature' not in st.session_state:
-        st.session_state.temperature = "0.8"
+Article: Murtagh a stickler for success Northeastern field hockey coach Cheryl Murtagh doesn't want the glare of the spotlight that shines on her to detract from a team that has been the America East champion for the past three years and has been to the NCAA tournament 13 times.
+Label:
+'''.strip()
+        st.session_state.temperature = "0.0"
+        st.session_state.top_p = "1.0"
+        st.session_state.max_new_tokens = "5"
+        st.session_state.stop = r'\n'
+        
+    elif st.session_state.preset == "Paraphrasing":
+        
+        st.session_state.prompt = '''
+Paraphrase the given sentence into a different sentence.
 
-    if 'top_p' not in st.session_state:
+Input: Can you recommed some upscale restaurants in New York?
+Output: What upscale restaurants do you recommend in New York?
+
+Input: What are the famous places we should not miss in Paris?
+Output: Recommend some of the best places to visit in Paris?
+
+Input: Could you recommed some hotels that have cheap price in Zurich?
+Output:
+'''.strip()
+        st.session_state.temperature = "0.8"
         st.session_state.top_p = "1.0"
+        st.session_state.max_new_tokens = "20"
+        st.session_state.stop = r'\n'
+        
+    elif st.session_state.preset == "Text Summarization":
+        
+        st.session_state.prompt = '''
+Given a review from Amazon's food products, the task is to generate a short summary of the given review in the input.
 
-    if 'top_k' not in st.session_state:
-        st.session_state.top_k = "40"
+Input: I have bought several of the Vitality canned dog food products and have found them all to be of good quality. The product looks more like a stew than a processed meat and it smells better. My Labrador is finicky and she appreciates this product better than most.
+Output: Good Quality Dog Food
+
+Input: Product arrived labeled as Jumbo Salted Peanuts...the peanuts were actually small sized unsalted. Not sure if this was an error or if the vendor intended to represent the product as 'Jumbo'.
+Output: Not as Advertised
+
+Input: Wolfgang Puck recently changed their Chicken soup. The old stuff is excellent. The new stuff is watered down, not filling and doesn't taste as good. The image and description on this item stated that it was the old soup. The soup I received is the new stuff. Being a food item, returns are not allowed. I feel this was false advertizing and am extremely displeased.
+Output:
+'''.strip()
+        st.session_state.temperature = "0.0"
+        st.session_state.top_p = "1.0"
+        st.session_state.max_new_tokens = "10"
+        st.session_state.stop = r'\n'
+        
+    elif st.session_state.preset == "Word Sense Disambiguation":
         
-    if 'max_new_tokens' not in st.session_state:
+        st.session_state.prompt = '''
+Identify which sense of a word is meant in a given context.
+
+Context: The river overflowed the bank.
+Word: bank
+Sense: river bank
+
+Context: A mouse takes much more room than a trackball.
+Word: mouse
+Sense: computer mouse
+
+Context: The bank will not be accepting cash on Saturdays.
+Word: bank
+Sense: commercial (finance) banks
+
+Context: Bill killed the project
+Word: kill
+Sense:
+'''.strip()
+        st.session_state.temperature = "0.0"
+        st.session_state.top_p = "1.0"
         st.session_state.max_new_tokens = "10"
+        st.session_state.stop = r'\n'
         
-    if 'updated' not in st.session_state:
-        st.session_state.updated = False
+    elif st.session_state.preset == "Natural Language Inference":
         
+        st.session_state.prompt = '''
+Given a pair of sentences, choose whether the two sentences agree (entailment)/disagree (contradiction) with each other.
+Possible labels: 1. entailment 2. contradiction
+
+Sentence 1: The skier was on the edge of the ramp. Sentence 2: The skier was dressed in winter clothes.
+Label: entailment
+
+Sentence 1: The boy skated down the staircase railing. Sentence 2: The boy is a newbie skater.
+Label: contradiction
+
+Sentence 1: Two middle-aged people stand by a golf hole. Sentence 2: A couple riding in a golf cart.
+Label:
+'''.strip()
+        st.session_state.temperature = "0.0"
+        st.session_state.top_p = "1.0"
+        st.session_state.max_new_tokens = "2"
+        st.session_state.stop = r'\n'
+    
+    else:
+        pass
+    
+    
+def main():
+
+    if 'preset' not in st.session_state:
+        st.session_state.preset = "Sentiment Analysis"
+        st.session_state.top_k = "40"
+        st.session_state.stop = r'\n'
+        set_preset()
 
     st.title("GPT-JT")
 
@@ -131,26 +233,27 @@ def main():
         top_p = st.text_input('top_p', st.session_state.top_p)
         # num_completions = st.text_input('num_completions (only the best one will be returend)', "1")
         num_completions = "1"
-        stop = st.text_input('stop, split by;', r'\n')
+        stop = st.text_input('stop, split by;', st.session_state.stop)
         # seed = st.text_input('seed', "42")
         seed = "42"
 
     with col2:
 
-        # preset = st.radio(
-        #     "Recommended Templates", 
-        #     ('Classification', 'Generation'), 
-        #     on_change=set_preset,
-        #     key="preset",
-        #     horizontal=True
-        # )
+        preset = st.selectbox(
+            label="Examples", 
+            options=('Question Answering', 'Sentiment Analysis', 
+                     "Topic Classification", "Paraphrasing", "Text Summarization",
+                     "Word Sense Disambiguation", "Natural Language Inference"), 
+            on_change=set_preset,
+            key="preset",
+        )
 
         prompt_area = st.empty()
         prompt = prompt_area.text_area(
             "Prompt",
             value=st.session_state.prompt,
             max_chars=4096,
-            height=300,
+            height=500,
         )
 
         generated_area = st.empty()