Update app.py

app.py

@@ -5,13 +5,12 @@ import requests
 
 #Streaming endpoint 
 API_URL = "https://api.openai.com/v1/chat/completions" #os.getenv("API_URL") + "/generate_stream"
-# OPENAI_API_KEY = os.getenv("ChatGPT") # Key 03-23
-
 OPENAI_API_KEY= os.environ["HF_TOKEN"] # Add a token to this space .  Then copy it to the repository secret in this spaces settings panel.  os.environ reads from there.
 # Keys for Open AI ChatGPT API usage are created from here: https://platform.openai.com/account/api-keys
 
 def predict(inputs, top_p, temperature, chat_counter, chatbot=[], history=[]):  #repetition_penalty, top_k
 
+    # 1. Set up a payload
     payload = {
     "model": "gpt-3.5-turbo",
     "messages": [{"role": "user", "content": f"{inputs}"}],
@@ -23,11 +22,13 @@ def predict(inputs, top_p, temperature, chat_counter, chatbot=[], history=[]):
     "frequency_penalty":0,
     }
 
+    # 2. Define your headers and add a key from https://platform.openai.com/account/api-keys
     headers = {
     "Content-Type": "application/json",
     "Authorization": f"Bearer {OPENAI_API_KEY}"
     }
 
+    # 3. Create a chat counter loop that feeds [Predict next best anything based on last input and attention with memory defined by introspective attention over time]
     print(f"chat_counter - {chat_counter}")
     if chat_counter != 0 :
         messages=[]
@@ -55,9 +56,9 @@ def predict(inputs, top_p, temperature, chat_counter, chatbot=[], history=[]):
         "presence_penalty":0,
         "frequency_penalty":0,
         }
-
     chat_counter+=1
 
+    # 4. POST it to OPENAI API
     history.append(inputs)
     print(f"payload is - {payload}")
     # make a POST request to the API endpoint using the requests.post method, passing in stream=True
@@ -66,6 +67,8 @@ def predict(inputs, top_p, temperature, chat_counter, chatbot=[], history=[]):
     token_counter = 0 
     partial_words = "" 
 
+    # 5. Iterate through response lines and structure readable response
+    # TODO - make this parse out markdown so we can have similar interface
     counter=0
     for chunk in response.iter_lines():
         #Skipping first chunk
@@ -93,18 +96,34 @@ def predict(inputs, top_p, temperature, chat_counter, chatbot=[], history=[]):
 def reset_textbox():
     return gr.update(value='')
 
-title = """<h1 align="center">🔥ChatGPT API 🚀Streaming🚀</h1>"""
-description = """Language models can be conditioned to act like dialogue agents through a conversational prompt that typically takes the form:
-```
-User: <utterance>
-Assistant: <utterance>
-User: <utterance>
-Assistant: <utterance>
-...
-```
-In this app, you can explore the outputs of a gpt-3.5-turbo LLM.
-"""
-                
+title = """<h1 align="center">Memory Chat Story Generator ChatGPT</h1>"""
+description = """
+
+## ChatGPT Datasets 📚
+- WebText
+- Common Crawl
+- BooksCorpus
+- English Wikipedia
+- Toronto Books Corpus
+- OpenWebText
+
+## ChatGPT Datasets - Details 📚
+- **WebText:** A dataset of web pages crawled from domains on the Alexa top 5,000 list. This dataset was used to pretrain GPT-2.
+  - [WebText: A Large-Scale Unsupervised Text Corpus by Radford et al.](https://paperswithcode.com/dataset/webtext)
+- **Common Crawl:** A dataset of web pages from a variety of domains, which is updated regularly. This dataset was used to pretrain GPT-3.
+  - [Language Models are Few-Shot Learners](https://paperswithcode.com/dataset/common-crawl) by Brown et al.
+- **BooksCorpus:** A dataset of over 11,000 books from a variety of genres.
+  - [Scalable Methods for 8 Billion Token Language Modeling](https://paperswithcode.com/dataset/bookcorpus) by Zhu et al.
+- **English Wikipedia:** A dump of the English-language Wikipedia as of 2018, with articles from 2001-2017.
+  - [Improving Language Understanding by Generative Pre-Training](https://huggingface.co/spaces/awacke1/WikipediaUltimateAISearch?logs=build) Space for Wikipedia Search
+- **Toronto Books Corpus:** A dataset of over 7,000 books from a variety of genres, collected by the University of Toronto.
+  - [Massively Multilingual Sentence Embeddings for Zero-Shot Cross-Lingual Transfer and Beyond](https://paperswithcode.com/dataset/bookcorpus) by Schwenk and Douze.
+- **OpenWebText:** A dataset of web pages that were filtered to remove content that was likely to be low-quality or spammy. This dataset was used to pretrain GPT-3.
+  - [Language Models are Few-Shot Learners](https://paperswithcode.com/dataset/openwebtext) by Brown et al.
+  
+  """
+
+# 6. Use Gradio to pull it all together
 with gr.Blocks(css = """#col_container {width: 1000px; margin-left: auto; margin-right: auto;}
                 #chatbot {height: 520px; overflow: auto;}""") as demo:
     gr.HTML(title)
@@ -115,12 +134,9 @@ with gr.Blocks(css = """#col_container {width: 1000px; margin-left: auto; margin
         state = gr.State([]) #s
         b1 = gr.Button()
     
-        #inputs, top_p, temperature, top_k, repetition_penalty
         with gr.Accordion("Parameters", open=False):
             top_p = gr.Slider( minimum=-0, maximum=1.0, value=1.0, step=0.05, interactive=True, label="Top-p (nucleus sampling)",)
             temperature = gr.Slider( minimum=-0, maximum=5.0, value=1.0, step=0.1, interactive=True, label="Temperature",)
-            #top_k = gr.Slider( minimum=1, maximum=50, value=4, step=1, interactive=True, label="Top-k",)
-            #repetition_penalty = gr.Slider( minimum=0.1, maximum=3.0, value=1.03, step=0.01, interactive=True, label="Repetition Penalty", )
             chat_counter = gr.Number(value=0, visible=False, precision=0)
 
     inputs.submit( predict, [inputs, top_p, temperature,chat_counter, chatbot, state], [chatbot, state, chat_counter],)
@@ -128,5 +144,5 @@ with gr.Blocks(css = """#col_container {width: 1000px; margin-left: auto; margin
     b1.click(reset_textbox, [], [inputs])
     inputs.submit(reset_textbox, [], [inputs])
                     
-    #gr.Markdown(description)
+    gr.Markdown(description)
     demo.queue().launch(debug=True)