Update app.py

app.py

@@ -1,63 +1,86 @@
 import gradio as gr
-from huggingface_hub import InferenceClient
+from gpt4all import GPT4All
+from huggingface_hub import hf_hub_download
+
+title = "DiarizationLM GGUF inference on CPU"
+
+description = """
 
 """
-For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
-"""
-client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
 
+model_path = "models"
+model_name = "model-unsloth.Q4_K_M.gguf"
+hf_hub_download(repo_id="google/DiarizationLM-13b-Fisher-v1", filename=model_name, local_dir=model_path, local_dir_use_symlinks=False)
 
-def respond(
-    message,
-    history: list[tuple[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-):
-    messages = [{"role": "system", "content": system_message}]
+print("Start the model init process")
+model = GPT4All(model_name, model_path, allow_download = False, device="cpu")
+print("Finish the model init process")
 
-    for val in history:
-        if val[0]:
-            messages.append({"role": "user", "content": val[0]})
-        if val[1]:
-            messages.append({"role": "assistant", "content": val[1]})
+model.config["promptTemplate"] = "{0} --> "
+model.config["systemPrompt"] = ""
+model._is_chat_session_activated = False
 
-    messages.append({"role": "user", "content": message})
+max_new_tokens = 2048
 
-    response = ""
+def generater(message, history, temperature, top_p, top_k):
+    prompt = model.config["promptTemplate"].format(message)
+    outputs = []    
+    for token in model.generate(prompt=prompt, temp=temperature, top_k = top_k, top_p = top_p, max_tokens = max_new_tokens, streaming=True):
+        outputs.append(token)
+        yield "".join(outputs)
 
-    for message in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
-        stream=True,
-        temperature=temperature,
-        top_p=top_p,
-    ):
-        token = message.choices[0].delta.content
+def vote(data: gr.LikeData):
+    if data.liked:
+        return
+    else:
+        return
 
-        response += token
-        yield response
+chatbot = gr.Chatbot(avatar_images=('resourse/user-icon.png', 'resourse/chatbot-icon.png'),bubble_full_width = False)
 
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
-demo = gr.ChatInterface(
-    respond,
-    additional_inputs=[
-        gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
-        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
-        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
-        gr.Slider(
-            minimum=0.1,
-            maximum=1.0,
-            value=0.95,
-            step=0.05,
-            label="Top-p (nucleus sampling)",
-        ),
-    ],
+additional_inputs=[
+    gr.Slider(
+        label="temperature",
+        value=0.5,
+        minimum=0.0,
+        maximum=2.0,
+        step=0.05,
+        interactive=True,
+        info="Higher values like 0.8 will make the output more random, while lower values like 0.2 will make it more focused and deterministic.",
+    ),
+    gr.Slider(
+        label="top_p",
+        value=1.0,
+        minimum=0.0,
+        maximum=1.0,
+        step=0.01,
+        interactive=True,
+        info="0.1 means only the tokens comprising the top 10% probability mass are considered. Suggest set to 1 and use temperature. 1 means 100% and will disable it",
+    ),
+    gr.Slider(
+        label="top_k",
+        value=40,
+        minimum=0,
+        maximum=1000,
+        step=1,
+        interactive=True,
+        info="limits candidate tokens to a fixed number after sorting by probability. Setting it higher than the vocabulary size deactivates this limit.",
+    )
+]
+
+iface = gr.ChatInterface(
+    fn = generater,
+    title=title,
+    description = description,
+    chatbot=chatbot,
+    additional_inputs=additional_inputs,
+    examples=[
+        ["<speaker:1> Hello, how are you doing <speaker:2> today? I am doing well."],
+   ]
 )
 
+with gr.Blocks(css="resourse/style/custom.css") as demo:
+    chatbot.like(vote, None, None)
+    iface.render()
 
 if __name__ == "__main__":
-    demo.launch()
+    demo.queue(max_size=3).launch()