Update app.py

app.py

@@ -1,41 +1,38 @@
 import gradio as gr
 from huggingface_hub import InferenceClient
 import random
-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]),
 ]
+
 def format_prompt(message, history):
     prompt = ""
     if history:
-        #<start_of_turn>userHow does the brain work?<end_of_turn><start_of_turn>model
         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}"
-    prompt += f"<start_of_turn>user{message}<end_of_turn><start_of_turn>model"
+            prompt += f"<start_of_turn>usuário{user_prompt}<end_of_turn>"
+            prompt += f"<start_of_turn>modelo{bot_response}"
+    prompt += f"<start_of_turn>usuário{message}<end_of_turn><start_of_turn>modelo"
     return prompt
 
-
-
-def chat_inf(system_prompt,prompt,history,client_choice,seed,temp,tokens,top_p,rep_p):
-    #token max=8192
-    client=clients[int(client_choice)-1]
+def chat_inf(system_prompt, prompt, history, client_choice, seed, temp, tokens, top_p, rep_p):
+    client = clients[int(client_choice) - 1]
     if not history:
         history = []
-        hist_len=0
+        hist_len = 0
     if history:
-        hist_len=len(history)
-        print(hist_len)
-        
-    #seed = random.randint(1,1111111111111111)
+        hist_len = len(history)
+
     generate_kwargs = dict(
         temperature=temp,
         max_new_tokens=tokens,
@@ -44,57 +41,56 @@ def chat_inf(system_prompt,prompt,history,client_choice,seed,temp,tokens,top_p,r
         do_sample=True,
         seed=seed,
     )
-    #formatted_prompt=prompt   
+    
     formatted_prompt = format_prompt(f"{system_prompt}, {prompt}", history)
     stream = client.text_generation(formatted_prompt, **generate_kwargs, stream=True, details=True, return_full_text=False)
     output = ""
-        
+    
     for response in stream:
         output += response.token.text
-        yield [(prompt,output)]
-    history.append((prompt,output))
+        yield [(prompt, output)]
+    history.append((prompt, output))
     yield history
 
 def clear_fn():
-    return None,None,None
-rand_val=random.randint(1,1111111111111111)
-def check_rand(inp,val):
-    if inp==True:
-        return gr.Slider(label="Seed", minimum=1, maximum=1111111111111111, value=random.randint(1,1111111111111111))
-    else:
-        return gr.Slider(label="Seed", minimum=1, maximum=1111111111111111, value=int(val))
+    return None, None, None
 
+rand_val = random.randint(1, 1111111111111111)
+
+def check_rand(inp, val):
+    if inp == True:
+        return gr.Slider(label="Semente", minimum=1, maximum=1111111111111111, value=random.randint(1, 1111111111111111))
+    else:
+        return gr.Slider(label="Semente", minimum=1, maximum=1111111111111111, value=int(val))
 
-    
 with gr.Blocks() as app:
-    gr.HTML("""<center><h1 style='font-size:xx-large;'>Google Gemma Models</h1><br><h3>running on Huggingface Inference Client</h3><br><h7>EXPERIMENTAL""")
+    gr.HTML("""<center><h1 style='font-size:xx-large;'>Modelos Google Gemma</h1><br><h3>Executando no Cliente de Inferência Huggingface</h3><br><h7>EXPERIMENTAL""")
     chat_b = gr.Chatbot(height=500)
     with gr.Group():
         with gr.Row():
             with gr.Column(scale=3):
                 inp = gr.Textbox(label="Prompt")
-                sys_inp = gr.Textbox(label="System Prompt (optional)")
+                sys_inp = gr.Textbox(label="Prompt do Sistema (opcional)")
                 with gr.Row():
                     with gr.Column(scale=2):
-                        btn = gr.Button("Chat")
+                        btn = gr.Button("Conversar")
                     with gr.Column(scale=1):
                         with gr.Group():
-                            stop_btn=gr.Button("Stop")
-                            clear_btn=gr.Button("Clear")                
-                client_choice=gr.Dropdown(label="Models",type='index',choices=[c for c in models],value=models[0],interactive=True)
+                            stop_btn = gr.Button("Parar")
+                            clear_btn = gr.Button("Limpar")
+                client_choice = gr.Dropdown(label="Modelos", type='index', choices=[c for c in models], value=models[0], interactive=True)
 
             with gr.Column(scale=1):
                 with gr.Group():
-                    rand = gr.Checkbox(label="Random Seed", value=True)
-                    seed=gr.Slider(label="Seed", minimum=1, maximum=1111111111111111,step=1, value=rand_val)
-                    tokens = gr.Slider(label="Max new tokens",value=6400,minimum=0,maximum=8000,step=64,interactive=True, visible=True,info="The maximum number of tokens")
-                    temp=gr.Slider(label="Temperature",step=0.01, minimum=0.01, maximum=1.0, value=0.9)
-                    top_p=gr.Slider(label="Top-P",step=0.01, minimum=0.01, maximum=1.0, value=0.9)
-                    rep_p=gr.Slider(label="Repetition Penalty",step=0.1, minimum=0.1, maximum=2.0, value=1.0)
-                    
+                    rand = gr.Checkbox(label="Semente Aleatória", value=True)
+                    seed = gr.Slider(label="Semente", minimum=1, maximum=1111111111111111, step=1, value=rand_val)
+                    tokens = gr.Slider(label="Máximo de novos tokens", value=6400, minimum=0, maximum=8000, step=64, interactive=True, visible=True, info="O número máximo de tokens")
+                    temp = gr.Slider(label="Temperatura", step=0.01, minimum=0.01, maximum=1.0, value=0.9)
+                    top_p = gr.Slider(label="Top-P", step=0.01, minimum=0.01, maximum=1.0, value=0.9)
+                    rep_p = gr.Slider(label="Penalidade de Repetição", step=0.1, minimum=0.1, maximum=2.0, value=1.0)
 
-    
-    go=btn.click(check_rand,[rand,seed],seed).then(chat_inf,[sys_inp,inp,chat_b,client_choice,seed,temp,tokens,top_p,rep_p],chat_b)
-    stop_btn.click(None,None,None,cancels=go)
-    clear_btn.click(clear_fn,None,[inp,sys_inp,chat_b])
-app.queue(default_concurrency_limit=10).launch()
+    go = btn.click(check_rand, [rand, seed], seed).then(chat_inf, [sys_inp, inp, chat_b, client_choice, seed, temp, tokens, top_p, rep_p], chat_b)
+    stop_btn.click(None, None, None, cancels=go)
+    clear_btn.click(clear_fn, None, [inp, sys_inp, chat_b])
+
+app.queue(default_concurrency_limit=10).launch()