Update app.py

app.py

@@ -23,7 +23,7 @@ if __name__ == "__main__":
     print(f"Number of threads available to the current process: {num_threads}")
 url = 'https://huggingface.co/TheBloke/Llama-2-7B-Chat-GGML/resolve/main/llama-2-7b-chat.ggmlv3.q2_K.bin'
 filename = wget.download(url)
-llm2 = Llama(model_path=filename, seed=random.randint(1, 2**31), lora_path="ggml-adapter-model.bin")
+llm2 = Llama(model_path=filename, seed=random.randint(1, 2**31))
 filename = wget.download(url)
 theme = gr.themes.Soft(
     primary_hue=gr.themes.Color("#ededed", "#fee2e2", "#fecaca", "#fca5a5", "#f87171", "#ef4444", "#dc2626", "#b91c1c", "#991b1b", "#7f1d1d", "#6c1e1e"),
@@ -33,17 +33,11 @@ title = """<h1 align="center">Chat with awesome LLAMA 2 CHAT model!</h1><br>"""
 with gr.Blocks(theme=theme) as demo:
     gr.HTML(title)
     gr.HTML("This model is awesome for its size! It is only 20th the size of Chatgpt but is still decent for chatting. However like all models, LLAMA-2-CHAT can hallucinate and provide incorrect information.")
-    chatbot = gr.Chatbot()
-    msg = gr.Textbox()
-    clear = gr.ClearButton([msg, chatbot])
+    #chatbot = gr.Chatbot()
+    #msg = gr.Textbox()
+    #clear = gr.ClearButton([msg, chatbot])
     #instruction = gr.Textbox(label="Instruction", placeholder=)
-
-    def user(user_message, history):
-        return gr.update(value="", interactive=True), history + [[user_message, None]]
-
-    def bot(history):
-        #instruction = history[-1][1] or ""
-        user_message = history[-1][0]
+    def bot(user_message):
         #token1 = llm.tokenize(b"### Instruction: ")
         #token2 = llm.tokenize(instruction.encode())
         #token3 = llm2.tokenize(b"USER: ")
@@ -53,21 +47,21 @@ with gr.Blocks(theme=theme) as demo:
         history[-1][1] = ""
         count = 0
         output = ""
+        outputs = ""
         for token in llm2.generate(tokens, top_k=50, top_p=0.73, temp=0.72, repeat_penalty=1.1):
             text = llm2.detokenize([token])
-            output += text.decode()
+            outputs += text.decode()
             count += 1
             if count >= 500 or (token == llm2.token_eos()):
                 break
-            history[-1][1] += text.decode()
-            yield history
-
-    response = msg.submit(user, [msg, chatbot], [msg, chatbot], queue=False).then(
-        bot, chatbot, chatbot
-    )
-    response.then(lambda: gr.update(interactive=True), None, [msg], queue=False)
+            output += text.decode()
+            yield output
     gr.HTML("Thanks for checking out this app!")
-
+    gr.Button("Answer").click(
+        fn=bot, 
+        inputs=gr.Textbox()
+        outputs=gr.Textbox()
+    )
 demo.queue()
 demo.launch(debug=True)