Update app.py

app.py

@@ -52,45 +52,45 @@ def format_prompt(message, history):
     print(prompt)
     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]
-    if not history:
-        history = []
-        hist_len=0
-    if history:
-        hist_len=len(history)
-        print(hist_len)
-    in_len=len(system_prompt+prompt)+hist_len
-    print("\n#########"+in_len)
-    if (in_len+tokens) > 8000:
-        yield [(prompt,"Wait. I need to compress our Chat history...")]
-        history=compress_history(history,client_choice,seed,temp,tokens,top_p,rep_p)
-        yield [(prompt,"History has been compressed, processing request...")]
+def chat_inf(system_prompt,prompt,history,client_choice,seed,temp,tokens,top_p,rep_p,hid_val):
+    if len(client_choice)>=hid_val:
             
-    generate_kwargs = dict(
-        temperature=temp,
-        max_new_tokens=tokens,
-        top_p=top_p,
-        repetition_penalty=rep_p,
-        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 = ""
+        #token max=8192
+        client=client_z[int(hid_val)-1]
         
-    for response in stream:
-        output += response.token.text
-        yield [(prompt,output)]
-    history.append((prompt,output))
-    yield history
-
+        if not history:
+            history = []
+            hist_len=0
+        if history:
+            hist_len=len(history)
+            print(hist_len)
+        in_len=len(system_prompt+prompt)+hist_len
+        print("\n#########"+in_len)
+        if (in_len+tokens) > 8000:
+            yield [(prompt,"Wait. I need to compress our Chat history...")]
+            #history=compress_history(history,client_choice,seed,temp,tokens,top_p,rep_p)
+            yield [(prompt,"History has been compressed, processing request...")]
+                
+        generate_kwargs = dict(
+            temperature=temp,
+            max_new_tokens=tokens,
+            top_p=top_p,
+            repetition_penalty=rep_p,
+            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 history
+    else:
+        yield None
 def clear_fn():
     return None,None,None
 rand_val=random.randint(1,1111111111111111)
@@ -120,7 +120,7 @@ with gr.Blocks() as app:
                         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],multiselect=True,interactive=True)
+                client_choice=gr.Dropdown(label="Models",type='index',choices=[c for c in models],max_choices=4,multiselect=True,interactive=True)
 
             with gr.Column(scale=1):
                 with gr.Group():
@@ -142,12 +142,20 @@ with gr.Blocks() as app:
                     wait_time=gr.Number(label="Wait Time",value=3000)
                     theme=gr.Radio(label="Theme", choices=["light","dark"],value="light")
                     chatblock=gr.Dropdown(label="Chatblocks",info="Choose specific blocks of chat",choices=[c for c in range(1,40)],multiselect=True)
-            
+    hid1=gr.Number(value=1)        
+    hid2=gr.Number(value=2)        
+    hid3=gr.Number(value=3)        
+    hid4=gr.Number(value=4)        
     client_choice.change(load_models,client_choice,[chat_a,chat_b,chat_c,chat_d])
 
     #im_go=im_btn.click(get_screenshot,[chat_b,im_height,im_width,chatblock,theme,wait_time],img)
     #chat_sub=inp.submit(check_rand,[rand,seed],seed).then(chat_inf,[sys_inp,inp,chat_b,client_choice,seed,temp,tokens,top_p,rep_p],chat_b)
-    #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)
+    
+    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,hid1],chat_b)
+    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,hid2],chat_b)
+    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,hid3],chat_b)
+    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,hid4],chat_b)
+    
     #stop_btn.click(None,None,None,cancels=[go,im_go,chat_sub])
     #clear_btn.click(clear_fn,None,[inp,sys_inp,chat_b])
 app.queue(default_concurrency_limit=10).launch()