Duplicate from yentinglin/Taiwan-LLaMa2

Co-authored-by: Yen-Ting Lin <yentinglin@users.noreply.huggingface.co>

.gitattributes

@@ -0,0 +1,35 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,13 @@
+---
+title: Tw Llama Demo
+emoji: 💻
+colorFrom: indigo
+colorTo: red
+sdk: gradio
+sdk_version: 3.39.0
+app_file: app.py
+pinned: false
+duplicated_from: yentinglin/Taiwan-LLaMa2
+---
+
+Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

app.py

@@ -0,0 +1,266 @@
+import os
+
+import gradio as gr
+from text_generation import Client
+from conversation import get_default_conv_template
+from transformers import AutoTokenizer
+from pymongo import MongoClient
+
+DB_NAME = os.getenv("MONGO_DBNAME", "taiwan-llm")
+USER = os.getenv("MONGO_USER")
+PASSWORD = os.getenv("MONGO_PASSWORD")
+
+uri = f"mongodb+srv://{USER}:{PASSWORD}@{DB_NAME}.kvwjiok.mongodb.net/?retryWrites=true&w=majority"
+mongo_client = MongoClient(uri)
+db = mongo_client[DB_NAME]
+conversations_collection = db['conversations']
+
+DESCRIPTION = """
+# Language Models for Taiwanese Culture
+
+<p align="center">
+✍️ <a href="https://huggingface.co/spaces/yentinglin/Taiwan-LLaMa2" target="_blank">Online Demo</a>  
+•
+🤗 <a href="https://huggingface.co/yentinglin" target="_blank">HF Repo</a> • 🐦 <a href="https://twitter.com/yentinglin56" target="_blank">Twitter</a> • 📃 <a href="https://arxiv.org/pdf/2305.13711.pdf" target="_blank">[Paper Coming Soon]</a>  
+• 👨️ <a href="https://github.com/MiuLab/Taiwan-LLaMa/tree/main" target="_blank">Github Repo</a> 
+    <br/><br/>
+    <img src="https://www.csie.ntu.edu.tw/~miulab/taiwan-llama/logo-v2.png" width="100"> <br/>
+</p>
+
+
+Taiwan-LLaMa is a fine-tuned model specifically designed for traditional mandarin applications. It is built upon the LLaMa 2 architecture and includes a pretraining phase with over 5 billion tokens and fine-tuning with over 490k multi-turn conversational data in Traditional Mandarin.
+
+## Key Features
+
+1. **Traditional Mandarin Support**: The model is fine-tuned to understand and generate text in Traditional Mandarin, making it suitable for Taiwanese culture and related applications.
+
+2. **Instruction-Tuned**: Further fine-tuned on conversational data to offer context-aware and instruction-following responses.
+
+3. **Performance on Vicuna Benchmark**: Taiwan-LLaMa's relative performance on Vicuna Benchmark is measured against models like GPT-4 and ChatGPT. It's particularly optimized for Taiwanese culture.
+
+4. **Flexible Customization**: Advanced options for controlling the model's behavior like system prompt, temperature, top-p, and top-k are available in the demo.
+
+## Model Versions
+
+Different versions of Taiwan-LLaMa are available:
+
+- **Taiwan-LLaMa v1.0 (This demo)**: Optimized for Taiwanese Culture
+- **Taiwan-LLaMa v0.9**: Partial instruction set
+- **Taiwan-LLaMa v0.0**: No Traditional Mandarin pretraining
+
+The models can be accessed from the provided links in the Hugging Face repository.
+
+Try out the demo to interact with Taiwan-LLaMa and experience its capabilities in handling Traditional Mandarin!
+"""
+
+LICENSE = """
+## Licenses
+
+- Code is licensed under Apache 2.0 License.
+- Models are licensed under the LLAMA 2 Community License.
+- By using this model, you agree to the terms and conditions specified in the license.
+- By using this demo, you agree to share your input utterances with us to improve the model.
+
+## Acknowledgements
+
+Taiwan-LLaMa project acknowledges the efforts of the [Meta LLaMa team](https://github.com/facebookresearch/llama) and [Vicuna team](https://github.com/lm-sys/FastChat) in democratizing large language models.
+"""
+
+DEFAULT_SYSTEM_PROMPT = "A chat between a curious user and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the user's questions. You are built by NTU Miulab by Yen-Ting Lin for research purpose."
+
+endpoint_url = os.environ.get("ENDPOINT_URL", "http://127.0.0.1:8080")
+client = Client(endpoint_url, timeout=120)
+eos_token = "</s>"
+MAX_MAX_NEW_TOKENS = 1024
+DEFAULT_MAX_NEW_TOKENS = 1024
+
+max_prompt_length = 4096 - MAX_MAX_NEW_TOKENS - 10
+
+model_name = "yentinglin/Taiwan-LLaMa-v1.0"
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+
+with gr.Blocks() as demo:
+    gr.Markdown(DESCRIPTION)
+
+    chatbot = gr.Chatbot()
+    with gr.Row():
+        msg = gr.Textbox(
+            container=False,
+            show_label=False,
+            placeholder='Type a message...',
+            scale=10,
+        )
+        submit_button = gr.Button('Submit',
+                                  variant='primary',
+                                  scale=1,
+                                  min_width=0)
+
+    with gr.Row():
+        retry_button = gr.Button('🔄  Retry', variant='secondary')
+        undo_button = gr.Button('↩️ Undo', variant='secondary')
+        clear = gr.Button('🗑️  Clear', variant='secondary')
+
+    saved_input = gr.State()
+
+    with gr.Accordion(label='Advanced options', open=False):
+        system_prompt = gr.Textbox(label='System prompt',
+                                   value=DEFAULT_SYSTEM_PROMPT,
+                                   lines=6)
+        max_new_tokens = gr.Slider(
+            label='Max new tokens',
+            minimum=1,
+            maximum=MAX_MAX_NEW_TOKENS,
+            step=1,
+            value=DEFAULT_MAX_NEW_TOKENS,
+        )
+        temperature = gr.Slider(
+            label='Temperature',
+            minimum=0.1,
+            maximum=1.0,
+            step=0.1,
+            value=0.7,
+        )
+        top_p = gr.Slider(
+            label='Top-p (nucleus sampling)',
+            minimum=0.05,
+            maximum=1.0,
+            step=0.05,
+            value=0.9,
+        )
+        top_k = gr.Slider(
+            label='Top-k',
+            minimum=1,
+            maximum=1000,
+            step=1,
+            value=50,
+        )
+
+    def user(user_message, history):
+        return "", history + [[user_message, None]]
+
+
+    def bot(history, max_new_tokens, temperature, top_p, top_k, system_prompt):
+        conv = get_default_conv_template("vicuna").copy()
+        roles = {"human": conv.roles[0], "gpt": conv.roles[1]}  # map human to USER and gpt to ASSISTANT
+        conv.system = system_prompt
+        for user, bot in history:
+            conv.append_message(roles['human'], user)
+            conv.append_message(roles["gpt"], bot)
+        msg = conv.get_prompt()
+        prompt_tokens = tokenizer.encode(msg)
+        length_of_prompt = len(prompt_tokens)
+        if length_of_prompt > max_prompt_length:
+            msg = tokenizer.decode(prompt_tokens[-max_prompt_length + 1:])
+
+        history[-1][1] = ""
+        for response in client.generate_stream(
+                msg,
+                max_new_tokens=max_new_tokens,
+                temperature=temperature,
+                top_p=top_p,
+                top_k=top_k,
+        ):
+            if not response.token.special:
+                character = response.token.text
+                history[-1][1] += character
+                yield history
+
+        # After generating the response, store the conversation history in MongoDB
+        conversation_document = {
+            "model_name": model_name,
+            "history": history,
+            "system_prompt": system_prompt,
+            "max_new_tokens": max_new_tokens,
+            "temperature": temperature,
+            "top_p": top_p,
+            "top_k": top_k,
+        }
+        conversations_collection.insert_one(conversation_document)
+
+    msg.submit(user, [msg, chatbot], [msg, chatbot], queue=False).then(
+        fn=bot,
+        inputs=[
+            chatbot,
+            max_new_tokens,
+            temperature,
+            top_p,
+            top_k,
+            system_prompt,
+        ],
+        outputs=chatbot
+    )
+    submit_button.click(
+        user, [msg, chatbot], [msg, chatbot], queue=False
+    ).then(
+        fn=bot,
+        inputs=[
+            chatbot,
+            max_new_tokens,
+            temperature,
+            top_p,
+            top_k,
+            system_prompt,
+        ],
+        outputs=chatbot
+    )
+
+
+    def delete_prev_fn(
+            history: list[tuple[str, str]]) -> tuple[list[tuple[str, str]], str]:
+        try:
+            message, _ = history.pop()
+        except IndexError:
+            message = ''
+        return history, message or ''
+
+
+    def display_input(message: str,
+                      history: list[tuple[str, str]]) -> list[tuple[str, str]]:
+        history.append((message, ''))
+        return history
+
+    retry_button.click(
+        fn=delete_prev_fn,
+        inputs=chatbot,
+        outputs=[chatbot, saved_input],
+        api_name=False,
+        queue=False,
+    ).then(
+        fn=display_input,
+        inputs=[saved_input, chatbot],
+        outputs=chatbot,
+        api_name=False,
+        queue=False,
+    ).then(
+        fn=bot,
+        inputs=[
+            chatbot,
+            max_new_tokens,
+            temperature,
+            top_p,
+            top_k,
+            system_prompt,
+        ],
+        outputs=chatbot,
+    )
+
+    undo_button.click(
+        fn=delete_prev_fn,
+        inputs=chatbot,
+        outputs=[chatbot, saved_input],
+        api_name=False,
+        queue=False,
+    ).then(
+        fn=lambda x: x,
+        inputs=[saved_input],
+        outputs=msg,
+        api_name=False,
+        queue=False,
+    )
+
+    clear.click(lambda: None, None, chatbot, queue=False)
+
+    gr.Markdown(LICENSE)
+
+demo.queue(concurrency_count=4, max_size=128)
+demo.launch()

conversation.py

@@ -0,0 +1,271 @@
+"""
+Conversation prompt template.
+Now we support
+- Vicuna
+- Koala
+- OpenAssistant/oasst-sft-1-pythia-12b
+- StabilityAI/stablelm-tuned-alpha-7b
+- databricks/dolly-v2-12b
+- THUDM/chatglm-6b
+- Alpaca/LLaMa
+"""
+
+import dataclasses
+from enum import auto, Enum
+from typing import List, Tuple, Any
+
+
+class SeparatorStyle(Enum):
+    """Different separator style."""
+
+    SINGLE = auto()
+    TWO = auto()
+    DOLLY = auto()
+    OASST_PYTHIA = auto()
+
+
+@dataclasses.dataclass
+class Conversation:
+    """A class that keeps all conversation history."""
+
+    system: str
+    roles: List[str]
+    messages: List[List[str]]
+    offset: int
+    sep_style: SeparatorStyle = SeparatorStyle.SINGLE
+    sep: str = "###"
+    sep2: str = None
+
+    # Used for gradio server
+    skip_next: bool = False
+    conv_id: Any = None
+
+    def get_prompt(self):
+        if self.sep_style == SeparatorStyle.SINGLE:
+            ret = self.system
+            for role, message in self.messages:
+                if message:
+                    ret += self.sep + " " + role + ": " + message
+                else:
+                    ret += self.sep + " " + role + ":"
+            return ret
+        elif self.sep_style == SeparatorStyle.TWO:
+            seps = [self.sep, self.sep2]
+            ret = self.system + seps[0]
+            for i, (role, message) in enumerate(self.messages):
+                if message:
+                    ret += role + ": " + message + seps[i % 2]
+                else:
+                    ret += role + ":"
+            return ret
+        elif self.sep_style == SeparatorStyle.DOLLY:
+            seps = [self.sep, self.sep2]
+            ret = self.system
+            for i, (role, message) in enumerate(self.messages):
+                if message:
+                    ret += role + ":\n" + message + seps[i % 2]
+                    if i % 2 == 1:
+                        ret += "\n\n"
+                else:
+                    ret += role + ":\n"
+            return ret
+        elif self.sep_style == SeparatorStyle.OASST_PYTHIA:
+            ret = self.system
+            for role, message in self.messages:
+                if message:
+                    ret += role + message + self.sep
+                else:
+                    ret += role
+            return ret
+        else:
+            raise ValueError(f"Invalid style: {self.sep_style}")
+
+    def append_message(self, role, message):
+        self.messages.append([role, message])
+
+    def to_gradio_chatbot(self):
+        ret = []
+        for i, (role, msg) in enumerate(self.messages[self.offset :]):
+            if i % 2 == 0:
+                ret.append([msg, None])
+            else:
+                ret[-1][-1] = msg
+        return ret
+
+    def copy(self):
+        return Conversation(
+            system=self.system,
+            roles=self.roles,
+            messages=[[x, y] for x, y in self.messages],
+            offset=self.offset,
+            sep_style=self.sep_style,
+            sep=self.sep,
+            sep2=self.sep2,
+            conv_id=self.conv_id,
+        )
+
+    def dict(self):
+        return {
+            "system": self.system,
+            "roles": self.roles,
+            "messages": self.messages,
+            "offset": self.offset,
+            "sep": self.sep,
+            "sep2": self.sep2,
+            "conv_id": self.conv_id,
+        }
+
+
+conv_one_shot = Conversation(
+    system="A chat between a curious human and an artificial intelligence assistant. "
+    "The assistant gives helpful, detailed, and polite answers to the human's questions.",
+    roles=("Human", "Assistant"),
+    messages=(
+        (
+            "Human",
+            "What are the key differences between renewable and non-renewable energy sources?",
+        ),
+        (
+            "Assistant",
+            "Renewable energy sources are those that can be replenished naturally in a relatively "
+            "short amount of time, such as solar, wind, hydro, geothermal, and biomass. "
+            "Non-renewable energy sources, on the other hand, are finite and will eventually be "
+            "depleted, such as coal, oil, and natural gas. Here are some key differences between "
+            "renewable and non-renewable energy sources:\n"
+            "1. Availability: Renewable energy sources are virtually inexhaustible, while non-renewable "
+            "energy sources are finite and will eventually run out.\n"
+            "2. Environmental impact: Renewable energy sources have a much lower environmental impact "
+            "than non-renewable sources, which can lead to air and water pollution, greenhouse gas emissions, "
+            "and other negative effects.\n"
+            "3. Cost: Renewable energy sources can be more expensive to initially set up, but they typically "
+            "have lower operational costs than non-renewable sources.\n"
+            "4. Reliability: Renewable energy sources are often more reliable and can be used in more remote "
+            "locations than non-renewable sources.\n"
+            "5. Flexibility: Renewable energy sources are often more flexible and can be adapted to different "
+            "situations and needs, while non-renewable sources are more rigid and inflexible.\n"
+            "6. Sustainability: Renewable energy sources are more sustainable over the long term, while "
+            "non-renewable sources are not, and their depletion can lead to economic and social instability.",
+        ),
+    ),
+    offset=2,
+    sep_style=SeparatorStyle.SINGLE,
+    sep="###",
+)
+
+
+conv_vicuna_v1_1 = Conversation(
+    system="A chat between a curious user and an artificial intelligence assistant. "
+    "The assistant gives helpful, detailed, and polite answers to the user's questions. You are built by NTU Miulab by Yen-Ting Lin for research purpose.",
+    # system="一位好奇的用戶和一個人工智能助理之間的聊天。你是一位助理。請對用戶的問題提供有用、詳細和有禮貌的答案。",
+    roles=("USER", "ASSISTANT"),
+    messages=(),
+    offset=0,
+    sep_style=SeparatorStyle.TWO,
+    sep=" ",
+    sep2="</s>",
+)
+
+conv_story = Conversation(
+    system="A chat between a curious user and an artificial intelligence assistant. "
+    "The assistant gives helpful, detailed, and polite answers to the user's questions.",
+    roles=("USER", "ASSISTANT"),
+    messages=(),
+    offset=0,
+    sep_style=SeparatorStyle.TWO,
+    sep=" ",
+    sep2="<|endoftext|>",
+)
+
+conv_koala_v1 = Conversation(
+    system="BEGINNING OF CONVERSATION:",
+    roles=("USER", "GPT"),
+    messages=(),
+    offset=0,
+    sep_style=SeparatorStyle.TWO,
+    sep=" ",
+    sep2="</s>",
+)
+
+conv_dolly = Conversation(
+    system="Below is an instruction that describes a task. Write a response that appropriately completes the request.\n\n",
+    roles=("### Instruction", "### Response"),
+    messages=(),
+    offset=0,
+    sep_style=SeparatorStyle.DOLLY,
+    sep="\n\n",
+    sep2="### End",
+)
+
+conv_oasst = Conversation(
+    system="",
+    roles=("<|prompter|>", "<|assistant|>"),
+    messages=(),
+    offset=0,
+    sep_style=SeparatorStyle.OASST_PYTHIA,
+    sep="<|endoftext|>",
+)
+
+conv_stablelm = Conversation(
+    system="""<|SYSTEM|># StableLM Tuned (Alpha version)
+- StableLM is a helpful and harmless open-source AI language model developed by StabilityAI.
+- StableLM is excited to be able to help the user, but will refuse to do anything that could be considered harmful to the user.
+- StableLM is more than just an information source, StableLM is also able to write poetry, short stories, and make jokes.
+- StableLM will refuse to participate in anything that could harm a human.
+""",
+    roles=("<|USER|>", "<|ASSISTANT|>"),
+    messages=(),
+    offset=0,
+    sep_style=SeparatorStyle.OASST_PYTHIA,
+    sep="",
+)
+
+conv_templates = {
+    "conv_one_shot": conv_one_shot,
+    "vicuna_v1.1": conv_vicuna_v1_1,
+    "koala_v1": conv_koala_v1,
+    "dolly": conv_dolly,
+    "oasst": conv_oasst,
+}
+
+
+def get_default_conv_template(model_name):
+    model_name = model_name.lower()
+    if "vicuna" in model_name or "output" in model_name:
+        return conv_vicuna_v1_1
+    elif "koala" in model_name:
+        return conv_koala_v1
+    elif "dolly-v2" in model_name:
+        return conv_dolly
+    elif "oasst" in model_name and "pythia" in model_name:
+        return conv_oasst
+    elif "stablelm" in model_name:
+        return conv_stablelm
+    return conv_one_shot
+
+
+def compute_skip_echo_len(model_name, conv, prompt):
+    model_name = model_name.lower()
+    if "chatglm" in model_name:
+        skip_echo_len = len(conv.messages[-2][1]) + 1
+    elif "dolly-v2" in model_name:
+        special_toks = ["### Instruction:", "### Response:", "### End"]
+        skip_echo_len = len(prompt)
+        for tok in special_toks:
+            skip_echo_len -= prompt.count(tok) * len(tok)
+    elif "oasst" in model_name and "pythia" in model_name:
+        special_toks = ["<|prompter|>", "<|assistant|>", "<|endoftext|>"]
+        skip_echo_len = len(prompt)
+        for tok in special_toks:
+            skip_echo_len -= prompt.count(tok) * len(tok)
+    elif "stablelm" in model_name:
+        special_toks = ["<|SYSTEM|>", "<|USER|>", "<|ASSISTANT|>"]
+        skip_echo_len = len(prompt)
+        for tok in special_toks:
+            skip_echo_len -= prompt.count(tok) * len(tok)
+    else:
+        skip_echo_len = len(prompt) + 1 - prompt.count("</s>") * 3
+    return skip_echo_len
+
+
+if __name__ == "__main__":
+    print(default_conversation.get_prompt())

requirements.txt

@@ -0,0 +1,3 @@
+text-generation==0.6.0
+transformers==4.31.0
+pymongo==4.4.1