app.py

import gradio as gr
from transformers import pipeline
import librosa

########################LLama model###############################
from transformers import AutoModelForCausalLM, AutoTokenizer

model_name_or_path = "TheBloke/llama2_7b_chat_uncensored-GPTQ"
# To use a different branch, change revision
# For example: revision="main"
model = AutoModelForCausalLM.from_pretrained(model_name_or_path,
                                             device_map="auto",
                                             trust_remote_code=True,
                                             revision="main",
                                             #quantization_config=QuantizationConfig(disable_exllama=True)
                                             )

tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
Llama_pipe = pipeline(
    "text-generation",
    model=model,
    tokenizer=tokenizer,
    max_new_tokens=40,
    do_sample=True,
    temperature=0.7,
    top_p=0.95,
    top_k=40,
    repetition_penalty=1.1
)
history="""User: Hello, Rally?
Rally: I'm happy to see you again. What you want to talk to day?
User: Let's talk about food
Rally: Sure.
User: I'm hungry right now. Do you know any Vietnamese food?"""

prompt_template = f"""<|im_start|>system
Write one sentence to continue the conversation<|im_end|>
{history}
Rally:"""
print(Llama_pipe(prompt_template)[0]['generated_text'])

def RallyRespone(chat_history, message):
    chat_history += "User: " + message + "\n"
    t_chat = Llama_pipe(prompt_template)[0]['generated_text']
    res = t_chat[t_chat.rfind("Rally: "):]
    return res

########################ASR model###############################

from transformers import WhisperProcessor, WhisperForConditionalGeneration

# load model and processor
processor = WhisperProcessor.from_pretrained("openai/whisper-base")
model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base")
model.config.forced_decoder_ids = None

sample_rate = 16000

def ASR_model(audio, sr=16000):
    DB_audio = audio
    input_features = processor(audio, sampling_rate=sr, return_tensors="pt").input_features
    # generate token ids
    predicted_ids = model.generate(input_features)
    # decode token ids to text
    transcription = processor.batch_decode(predicted_ids, skip_special_tokens=False)

    transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True)

    return transcription



########################Gradio UI###############################
    
# Chatbot demo with multimodal input (text, markdown, LaTeX, code blocks, image, audio, & video). Plus shows support for streaming text.
def add_file(files):
    return files.name

def print_like_dislike(x: gr.LikeData):
    print(x.index, x.value, x.liked)

def upfile(files):
    x = librosa.load(files, sr=16000)
    print(x[0])
    text = ASR_model(x[0])
    return [text[0], text[0]]

def transcribe(audio):
    sr, y = audio
    y = y.astype(np.float32)
    y /= np.max(np.abs(y))

    return transcriber({"sampling_rate": sr, "raw": y})["text"], transcriber({"sampling_rate": sr, "raw": y})["text"]


# def recommand(text):
#     ret = "answer for"

#     return ret + text

def add_text(history, text):
    history = history + [(text, None)]
    return history, gr.Textbox(value="", interactive=False)

# def bot(history):
#     response = "**That's cool!**"
#     history[-1][1] = ""
#     for character in response:
#         history[-1][1] += character
#         time.sleep(0.05)
#         yield history

with gr.Blocks() as demo:
    chatbot = gr.Chatbot(
        [],
        elem_id="chatbot",
        bubble_full_width=False,
    )
    file_output = gr.File()

    def respond(message, chat_history):
        bot_message = RallyRespone(chat_history, message)


        chat_history.append((message, bot_message))
        time.sleep(2)
        print (chat_history[-1])
        return chat_history[-1][-1], chat_history

    with gr.Row():
        with gr.Column():
            audio_speech = gr.Audio(sources=["microphone"])
            submit = gr.Button("Submit")
            send = gr.Button("Send")
            btn = gr.UploadButton("📁", file_types=["audio"])

        with gr.Column():
            opt1 = gr.Button("1: ")
            opt2 = gr.Button("2: ")

    #submit.click(translate, inputs=audio_speech, outputs=[opt1, opt2])
    # output is opt1 value, opt2 value [ , ]

    file_msg = btn.upload(add_file, btn, file_output)
    submit.click(upfile, inputs=file_output, outputs=[opt1, opt2])
    send.click(transcribe, inputs=audio_speech, outputs=[opt1, opt2])
    opt1.click(respond, [opt1, chatbot], [opt1, chatbot])

    opt2.click(respond, [opt2, chatbot], [opt2, chatbot])

    #opt2.click(recommand, inputs=opt2)
    #click event maybe BOT . generate history = optx.value,

    chatbot.like(print_like_dislike, None, None)

if __name__ == "__main__":
    demo.queue()
    demo.launch(debug=True)