Update app.py

app.py

@@ -4,27 +4,25 @@ from huggingface_hub import InferenceClient
 
 ss_client = Client("https://omnibus-html-image-current-tab.hf.space/")
 
-models=[
+models = [
     "google/gemma-7b",
     "google/gemma-7b-it",
     "google/gemma-2b",
     "google/gemma-2b-it"
 ]
-clients=[
-InferenceClient(models[0]),
-InferenceClient(models[1]),
-InferenceClient(models[2]),
-InferenceClient(models[3]),
+clients = [
+    InferenceClient(models[0]),
+    InferenceClient(models[1]),
+    InferenceClient(models[2]),
+    InferenceClient(models[3]),
 ]
 
-VERBOSE=False
+VERBOSE = False
+
+
+def load_models():
+    return gr.update(label=models[0])
 
-def load_models(inp):
-    if VERBOSE==True:    
-        print(type(inp))
-        print(inp)
-        print(models[inp])
-    return gr.update(label=models[inp])
 
 def format_prompt(message, history):
     prompt = ""
@@ -32,28 +30,34 @@ def format_prompt(message, history):
         for user_prompt, bot_response in history:
             prompt += f"<start_of_turn>user{user_prompt}<end_of_turn>"
             prompt += f"<start_of_turn>model{bot_response}<end_of_turn>"
-            if VERBOSE==True:
+            if VERBOSE:
                 print(prompt)
     prompt += message
     return prompt
 
-def chat_inf(prompt,history,memory,client_choice,temp,tokens,top_p,rep_p,chat_mem):
-    hist_len=0
-    client=clients[int(client_choice)-1]
+
+def chat_inf(prompt, history, memory, temp, tokens, top_p, rep_p, chat_mem):
+    hist_len = 0
+    client = clients[0]
     if not history:
         history = []
-        hist_len=0
+        hist_len = 0
     if not memory:
         memory = []
-        mem_len=0        
+        mem_len = 0
     if memory:
-        for ea in memory[0-chat_mem:]:
-            hist_len+=len(str(ea))
-    in_len=len(prompt)+hist_len
+        for ea in memory[0 - chat_mem :]:
+            hist_len += len(str(ea))
+    in_len = len(prompt) + hist_len
 
-    if (in_len+tokens) > 8000:
-        history.append((prompt,"Wait, that's too many tokens, please reduce the 'Chat Memory' value, or reduce the 'Max new tokens' value"))
-        yield history,memory
+    if (in_len + tokens) > 8000:
+        history.append(
+            (
+                prompt,
+                "Wait, that's too many tokens, please reduce the 'Chat Memory' value, or reduce the 'Max new tokens' value",
+            )
+        )
+        yield history, memory
     else:
         generate_kwargs = dict(
             temperature=temp,
@@ -62,33 +66,59 @@ def chat_inf(prompt,history,memory,client_choice,temp,tokens,top_p,rep_p,chat_me
             repetition_penalty=rep_p,
             do_sample=True,
         )
-        formatted_prompt = format_prompt(prompt, memory[0-chat_mem:])
-        stream = client.text_generation(formatted_prompt, **generate_kwargs, stream=True, details=True, return_full_text=True)
+        formatted_prompt = format_prompt(prompt, memory[0 - chat_mem :])
+        stream = client.text_generation(
+            formatted_prompt,
+            **generate_kwargs,
+            stream=True,
+            details=True,
+            return_full_text=True,
+        )
         output = ""
         for response in stream:
             output += response.token.text
-            yield [(prompt,output)],memory
-        history.append((prompt,output))
-        memory.append((prompt,output))
-        yield history,memory
-        
-    if VERBOSE==True:
-        print("\n######### HIST "+str(in_len))
-        print("\n######### TOKENS "+str(tokens))        
-
-def get_screenshot(chat: list,height=5000,width=600,chatblock=[],theme="light",wait=3000,header=True):
+            yield [(prompt, output)], memory
+        history.append((prompt, output))
+        memory.append((prompt, output))
+        yield history, memory
+
+    if VERBOSE:
+        print("\n######### HIST " + str(in_len))
+        print("\n######### TOKENS " + str(tokens))
+
+
+def get_screenshot(
+    chat: list,
+    height=5000,
+    width=600,
+    chatblock=[],
+    theme="light",
+    wait=3000,
+    header=True,
+):
     tog = 0
     if chatblock:
         tog = 3
-    result = ss_client.predict(str(chat),height,width,chatblock,header,theme,wait,api_name="/run_script")
+    result = ss_client.predict(
+        str(chat),
+        height,
+        width,
+        chatblock,
+        header,
+        theme,
+        wait,
+        api_name="/run_script",
+    )
     out = f'https://omnibus-html-image-current-tab.hf.space/file={result[tog]}'
     return out
 
+
 def clear_fn():
-    return None,None,None,None
+    return None, None, None, None
+
 
 with gr.Blocks() as app:
-    memory=gr.State()
+    memory = gr.State()
     chat_b = gr.Chatbot(height=500)
     with gr.Group():
         with gr.Row():
@@ -97,17 +127,46 @@ with gr.Blocks() as app:
                 btn = gr.Button("Chat")
             with gr.Column(scale=1):
                 with gr.Group():
-                    temp=gr.Slider(label="Temperature",step=0.01, minimum=0.01, maximum=1.0, value=0.49)
-                    tokens = gr.Slider(label="Max new tokens",value=1600,minimum=0,maximum=8000,step=64,interactive=True, visible=True,info="The maximum number of tokens")
-                    top_p=gr.Slider(label="Top-P",step=0.01, minimum=0.01, maximum=1.0, value=0.49)
-                    rep_p=gr.Slider(label="Repetition Penalty",step=0.01, minimum=0.1, maximum=2.0, value=0.99)
-                    chat_mem=gr.Number(label="Chat Memory", info="Number of previous chats to retain",value=4)
-
-    client_choice=gr.Dropdown(label="Models",type='index',choices=[c for c in models],value=models[0],interactive=True)
-    client_choice.change(load_models,client_choice,[chat_b])
-    app.load(load_models,client_choice,[chat_b])
-    
-    chat_sub=inp.submit().then(chat_inf,[inp,chat_b,memory,client_choice,temp,tokens,top_p,rep_p,chat_mem],[chat_b,memory])
-    go=btn.click().then(chat_inf,[inp,chat_b,memory,client_choice,temp,tokens,top_p,rep_p,chat_mem],[chat_b,memory])
-
-    app.queue(default_concurrency_limit=10).launch()
+                    temp = gr.Slider(
+                        label="Temperature",
+                        step=0.01,
+                        minimum=0.01,
+                        maximum=1.0,
+                        value=0.49,
+                    )
+                    tokens = gr.Slider(
+                        label="Max new tokens",
+                        value=1600,
+                        minimum=0,
+                        maximum=8000,
+                        step=64,
+                        interactive=True,
+                        visible=True,
+                        info="The maximum number of tokens",
+                    )
+                    top_p = gr.Slider(
+                        label="Top-P",
+                        step=0.01,
+                        minimum=0.01,
+                        maximum=1.0,
+                        value=0.49,
+                    )
+                    rep_p = gr.Slider(
+                        label="Repetition Penalty",
+                        step=0.01,
+                        minimum=0.1,
+                        maximum=2.0,
+                        value=0.99,
+                    )
+                    chat_mem = gr.Number(
+                        label="Chat Memory",
+                        info="Number of previous chats to retain",
+                        value=4,
+                    )
+
+    app.load(load_models)
+    chat_sub = inp.submit().then(
+        chat_inf, [inp, chat_b, memory, temp, tokens, top_p, rep_p, chat_mem], [chat_b, memory]
+    )
+    go = btn.click().then(
+        chat_inf,