test new version mixtral-api

app.py

@@ -1,54 +1,56 @@
+import spaces
 from huggingface_hub import InferenceClient
 import gradio as gr
-import torch
-import spaces
-
-client = InferenceClient("mistralai/Mixtral-8x7B-Instruct-v0.1")
-
-def format_prompt(message, history):
-    prompt = "<s>"
-    for user_prompt, bot_response in history:
-        prompt += f"[INST] {user_prompt} [/INST]"
-        prompt += f" {bot_response}</s> "
-    prompt += f"[INST] {message} [/INST]"
-    return prompt
-
-@spaces.GPU
-def generate(
-    prompt, history, temperature=0, max_new_tokens=3500, top_p=0.95, repetition_penalty=1.0,
-):
-    temperature = float(temperature)
-    if temperature < 1e-2:
-        temperature = 1e-2
-    top_p = float(top_p)
-
-    generate_kwargs = dict(
-        temperature=temperature,
-        max_new_tokens=max_new_tokens,
-        top_p=top_p,
-        repetition_penalty=repetition_penalty,
-        do_sample=True,
-        seed=42,
-    )
-
-    formatted_prompt = format_prompt(prompt, history)
 
-    stream = client.text_generation(formatted_prompt, **generate_kwargs, stream=True, details=True, return_full_text=False)
-    output = ""
+client = InferenceClient('mistralai/Mixtral-8x7B-Instruct-v0.1')
 
+@spaces.GPU(duration=60)
+def generate_response(chat, kwargs):
+    output = ''
+    stream = client.text_generation(chat, **kwargs, stream=True, details=True, return_full_text=False)
     for response in stream:
         output += response.token.text
-        yield output
     return output
 
-mychatbot = gr.Chatbot(
-    bubble_full_width=False, show_label=False, show_copy_button=True, likeable=True,)
-
-demo = gr.ChatInterface(fn=generate,
-                        chatbot=mychatbot,
-                        title="Test API :)",
-                        retry_btn=None,
-                        undo_btn=None
-                       )
+def function(prompt, history):
+    chat = "<s>"
+    for user_prompt, bot_response in history:
+        chat += f"[INST] {user_prompt} [/INST] {bot_response}</s> "
+    chat += f"[INST] {prompt} [/INST]"
+    kwargs = dict(
+        temperature=0.80,
+        max_new_tokens=2048,
+        top_p=0.95,
+        repetition_penalty=1.0,
+        do_sample=True,  # Upewnij się, że używasz próbkowania
+        seed=1337
+    )
 
-demo.queue().launch(show_api=True)
+    try:
+        output = generate_response(chat, kwargs)
+        yield output
+    except:
+        yield ''
+
+interface = gr.ChatInterface(
+    fn=function,
+    chatbot=gr.Chatbot(
+        avatar_images=None,
+        container=False,
+        show_copy_button=True,
+        layout='bubble',
+        render_markdown=True,
+        line_breaks=True
+    ),
+    css='h1 {font-size:22px;} h2 {font-size:20px;} h3 {font-size:18px;} h4 {font-size:16px;}',
+    autofocus=True,
+    fill_height=True,
+    analytics_enabled=False,
+    submit_btn='Chat',
+    stop_btn=None,
+    retry_btn=None,
+    undo_btn=None,
+    clear_btn=None
+)
+
+interface.launch()

app.py_OLD

@@ -0,0 +1,54 @@
+from huggingface_hub import InferenceClient
+import gradio as gr
+import torch
+import spaces
+
+client = InferenceClient("mistralai/Mixtral-8x7B-Instruct-v0.1")
+
+def format_prompt(message, history):
+    prompt = "<s>"
+    for user_prompt, bot_response in history:
+        prompt += f"[INST] {user_prompt} [/INST]"
+        prompt += f" {bot_response}</s> "
+    prompt += f"[INST] {message} [/INST]"
+    return prompt
+
+@spaces.GPU
+def generate(
+    prompt, history, temperature=0, max_new_tokens=3500, top_p=0.95, repetition_penalty=1.0,
+):
+    temperature = float(temperature)
+    if temperature < 1e-2:
+        temperature = 1e-2
+    top_p = float(top_p)
+
+    generate_kwargs = dict(
+        temperature=temperature,
+        max_new_tokens=max_new_tokens,
+        top_p=top_p,
+        repetition_penalty=repetition_penalty,
+        do_sample=True,
+        seed=42,
+    )
+
+    formatted_prompt = format_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 output
+    return output
+
+mychatbot = gr.Chatbot(
+    bubble_full_width=False, show_label=False, show_copy_button=True, likeable=True,)
+
+demo = gr.ChatInterface(fn=generate,
+                        chatbot=mychatbot,
+                        title="Test API :)",
+                        retry_btn=None,
+                        undo_btn=None
+                       )
+
+demo.queue().launch(show_api=True)