Update app.py

app.py

@@ -2,22 +2,15 @@ import gradio as gr
 from gpt4all import GPT4All
 from huggingface_hub import hf_hub_download
 
-title = "Mistral-7B-Instruct-GGUF Run On CPU-Basic Free Hardware"
+title = "DiarizationLM GGUF inference on CPU"
 
 description = """
-🔎 [Mistral AI's Mistral 7B Instruct v0.1](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.1) [GGUF format model](https://huggingface.co/TheBloke/Mistral-7B-Instruct-v0.1-GGUF) , 4-bit quantization balanced quality gguf version, running on CPU. English Only (Also support other languages but the quality's not good). Using [GitHub - llama.cpp](https://github.com/ggerganov/llama.cpp) [GitHub - gpt4all](https://github.com/nomic-ai/gpt4all). 
-
-🔨 Running on CPU-Basic free hardware. Suggest duplicating this space to run without a queue. 
-
-Mistral does not support system prompt symbol (such as ```<<SYS>>```) now, input your system prompt in the first message if you need. Learn more: [Guardrailing Mistral 7B](https://docs.mistral.ai/usage/guardrailing). 
-"""
-
-"""
-[Model From TheBloke/Mistral-7B-Instruct-v0.1-GGUF](https://huggingface.co/TheBloke/Mistral-7B-Instruct-v0.1-GGUF)
-[Mistral-instruct-v0.1 System prompt](https://docs.mistral.ai/usage/guardrailing)
+DiarizationLM GGUF inference on CPU
 """
 
 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)
 model_name = "mistral-7b-instruct-v0.1.Q4_K_M.gguf"
 hf_hub_download(repo_id="TheBloke/Mistral-7B-Instruct-v0.1-GGUF", filename=model_name, local_dir=model_path, local_dir_use_symlinks=False)
 
@@ -25,35 +18,37 @@ print("Start the model init process")
 model = model = GPT4All(model_name, model_path, allow_download = False, device="cpu")
 print("Finish the model init process")
 
-model.config["promptTemplate"] = "[INST] {0} [/INST]"
+model.config["promptTemplate"] = "{0} --> "
 model.config["systemPrompt"] = ""
 model._is_chat_session_activated = False
 
 max_new_tokens = 2048
 
+print("Finish the model config process")
+
 def generater(message, history, temperature, top_p, top_k):
-    prompt = "<s>"
-    for user_message, assistant_message in history:
-        prompt += model.config["promptTemplate"].format(user_message)
-        prompt += assistant_message + "</s>"
-    prompt += model.config["promptTemplate"].format(message)
+    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)
 
+
 def vote(data: gr.LikeData):
     if data.liked:
         return
     else:
         return
 
-chatbot = gr.Chatbot(avatar_images=('resourse/user-icon.png', 'resourse/chatbot-icon.png'),bubble_full_width = False)
+print("Create chatbot")
+chatbot = gr.Chatbot()
+print("Created chatbot")
 
+print("Add additional_inputs")
 additional_inputs=[
     gr.Slider(
         label="temperature",
-        value=0.5,
+        value=0.0,
         minimum=0.0,
         maximum=2.0,
         step=0.05,
@@ -71,7 +66,7 @@ additional_inputs=[
     ),
     gr.Slider(
         label="top_k",
-        value=40,
+        value=50,
         minimum=0,
         maximum=1000,
         step=1,
@@ -79,9 +74,7 @@ additional_inputs=[
         info="limits candidate tokens to a fixed number after sorting by probability. Setting it higher than the vocabulary size deactivates this limit.",
     )
 ]
-
-character = "Sherlock Holmes"
-series = "Arthur Conan Doyle's novel"
+print("Added additional_inputs")
 
 iface = gr.ChatInterface(
     fn = generater,
@@ -90,17 +83,17 @@ iface = gr.ChatInterface(
     chatbot=chatbot,
     additional_inputs=additional_inputs,
     examples=[
-        ["Hello there! How are you doing?"],
-        ["How many hours does it take a man to eat a Helicopter?"],
-        ["You are a helpful and honest assistant. Always answer as helpfully as possible. If a question does not make any sense, or is not factually coherent, explain why instead of answering something not correct. If you don't know the answer to a question, please don't share false information."],
-        ["I want you to act as a spoken English teacher and improver. I will speak to you in English and you will reply to me in English to practice my spoken English. I want you to strictly correct my grammar mistakes, typos, and factual errors. I want you to ask me a question in your reply. Now let's start practicing, you could ask me a question first. Remember, I want you to strictly correct my grammar mistakes, typos, and factual errors."],
-        [f"I want you to act like {character} from {series}. I want you to respond and answer like {character} using the tone, manner and vocabulary {character} would use. Do not write any explanations. Only answer like {character}. You must know all of the knowledge of {character}."]
-    ]
+        ["<speaker:1> Hello, how are you doing <speaker:2> today? I am doing well."],
+   ]
 )
 
-with gr.Blocks(css="resourse/style/custom.css") as demo:
+print("Added iface")
+
+with gr.Blocks() as demo:
     chatbot.like(vote, None, None)
     iface.render()
 
+print("Rendered iface")
+
 if __name__ == "__main__":
     demo.queue(max_size=3).launch()