chat.py

import json
import re
import time
from datetime import datetime
import gradio as gr

import chat_client

CHAT_URL = "wss://chat.petals.dev/api/v2/generate"
#CHAT_URL='ws://localhost:8000/api/v2/generate'

EMPTY_STATE = {
    "generate": False,
    "model": None,
    "client": None,
    "history": [],
}


def generate(state, prompt, model, context, output, *args):
    # Save that we're in generating loop
    state["generate"] = True

    try:
        yield from _generate(state, prompt, model, context, output, *args)
    except (json.decoder.JSONDecodeError, BrokenPipeError):
        # Broken session, try to renew
        # TODO This is a bit fragile because of recursive call...
        print("Retrying session...")
        context = output
        output = ""
        yield from generate(state, prompt, model, context, output, *args)
    finally:
        state["generate"] = False


def _generate(
    state,
    prompt,
    model,
    context,
    output,
    endseq,
    max_length,
    do_sample,
    top_k,
    top_p,
    temperature,
):

    start = time.time()
    cnt = 0  # Tokens generated

    def stats():
        # Produces inline stats for generation speed
        if cnt == 0:
            return "\u2026 | ? sec/t"
        if cnt > time.time() - start:
            items_per_sec = cnt / (time.time() - start)
            return f" | {items_per_sec:.1f} t/sec"
        sec_per_item = (time.time() - start) / cnt
        return f" | {sec_per_item:.1f} sec/t"

    eos = "</s>\n" if "bloomz" in model else "\n\n"

    if state["model"] != model and output:
        # If the connection is resumed, output is truncated in generate().
        # So this executes when user change model.
        context = output
        output = ""

    # Update widgets even before we get the first response
    print("prompt", prompt)
    yield state, state["history"] + [[prompt, stats()]], "", output

    if (
        state["model"] != model
        or state["client"] == None
        or state["client"].is_session() == False
    ):

        try:
            state["client"] = chat_client.ModelClient(CHAT_URL)
            state["client"].open_session(model, max_length)
            state["model"] = model
        except Exception as e:
            print(datetime.now(), str(e)[-500:])
            raise gr.Error(str(e)[-500:])

    else:
        context = ""

    client = state["client"]
    context += eos

    # Fix eventual eos token mismatch and add eos token to context and prompt
    if "bloomz" in model:
        context = context.replace("\n\n", eos)
        prompt2 = prompt.replace("\n\n", eos) + "</s>\n"
    else:
        context = context.replace("</s>", eos)
        context = re.sub(r"\n\n+", "\n\n", context)
        prompt2 = prompt.replace("</s>", eos) + "\n\n"

    prompt2 = f"{context}Human: {prompt2}AI:"

    # Translate checkbox items to actual sequences
    seq = []
    for s in endseq:
        if s == "Human:":
            seq.append("Human:")
        if s == "AI:":
            seq.append("AI:")
        if s == "\\n":
            seq.append("\n")
        elif s == "</s>":
            seq.append("</s>")
        elif s == "? (question mark)":
            seq.append("?")
        elif s == ". (dot)":
            seq.append(".")

    # only top_k or top_p can be set
    if top_k == 0:
        top_k = None
    if top_p == 0:
        top_p = None
    if top_p and top_k:
        top_k = None

    if temperature == 0:
        temperature = 1.0

    output += prompt2

    orig_history = state["history"]
    new_line = ""
    try:
        for out in client.generate(
            prompt2,
            max_new_tokens=1,
            do_sample=do_sample,
            temperature=temperature,
            top_k=top_k,
            top_p=top_p,
            stop_sequences=seq,
        ):

            if not state["generate"]:
                client.close_session()
                yield state, [], "", ""
                # Stopping generation
                return

            cnt += 1
            new_line += out

            # Detect end sequences and finish the generation
            # prematurely if found.
            for s in seq:
                spl = new_line.split(s)
                new_line = spl[0]
                if len(spl) > 1:
                    state["history"] = orig_history + [[prompt, new_line]]
                    output += new_line
                    yield state, state["history"], "", output
                    # Stopping generation
                    return

            # Keep original history untouched as we're adding just
            # a chunks at one moment.
            state["history"] = orig_history + [[prompt, new_line + stats()]]
            yield state, state["history"], "", output

            # Avoid throwing an exception by generate()
            # to prevent UI errors.
            if cnt >= max_length - 6: # FIXME Bulgarian constant
                break

        # Final line w/o statistics
        yield state, state["history"], "", output

    except (json.decoder.JSONDecodeError, BrokenPipeError):
        # Session was interrupted
        # Handled in upstream func
        client.close_session()
        state["client"] = None
        state["model"] = None

        print("Broken session!")
        raise
    except Exception as e:
        client.close_session()
        state["client"] = None
        state["model"] = None

        print(datetime.now(), str(e)[-500:])
        raise gr.Error(str(e)[-500:])


def reset(state):
    """Resets the session and clears the chat window."""
    state.update(EMPTY_STATE)
    return state, [], ""


# ---------------------------------------------------------
# Defining Gradio layout
with gr.Blocks() as iface_chat:
    gr.Markdown("""**Let's talk to AI in a chat!**""")

    with gr.Row():
        model = gr.Radio(
            ["stabilityai/StableBeluga2", "meta-llama/Llama-2-70b-chat-hf", "bigscience/bloomz"], value="stabilityai/StableBeluga2", label="Use model"
        )

        # Additional ending sequence, at which generation shoud stop
        endseq = gr.CheckboxGroup(
            ["Human:", "AI:", "\\n", "</s>", "? (question mark)", ". (dot)"],
            value=["Human:", "AI:", "</s>"],
            label="Extra end sequences",
        )

        # Maximum length of inference session
        max_length = gr.Radio(
            [64, 128, 256, 512, 1024, 2048],
            value=1024,
            interactive=True,
            label="Max length",
        )

    with gr.Row():
        with gr.Column():
            # Switch between sampling and greedy generation
            do_sample = gr.Checkbox(value=True, interactive=True, label="do_sample")
            context = gr.Textbox(
                lines=3,
                label="Initial context:",
                interactive=True,
                value="A Human talks to a powerful AI that follows "
                "the Human's instructions.\n"
                "AI is talkative, friendly, positive and provides "
                "detailed answers to any question.</s>\n"
                "Human: Hi!</s>\n"
                "AI: How can I help you?",
            )

        # Only one of top_k and top_p can be set. Requires "do_sample=True" to work.
        top_k = gr.Number(value=0, precision=0, interactive=True, label="top_k")
        top_p = gr.Number(value=0.9, precision=2, interactive=True, label="top_p")
        # TODO num_beams

        # Generation temperature
        temperature = gr.Number(
            value=0.75, precision=2, interactive=True, label="Temperature"
        )

    chat = gr.Chatbot(label="Chat window")
    prompt = gr.Textbox(
        show_label=False, label="Prompt", placeholder="Prompt Here and press Enter..."
    ).style(container=False)

    with gr.Row():
        button_generate = gr.Button("Generate")
        button_reset = gr.Button("Reset session")

    with gr.Accordion("Raw prompt log", open=False):
        output = gr.Textbox(lines=3, show_label=False).style(container=False)

    # Chat history
    state = gr.State(EMPTY_STATE)

    # Define button actions
    inputs = [
        state,
        prompt,
        model,
        context,
        output,
        endseq,
        max_length,
        do_sample,
        top_k,
        top_p,
        temperature,
    ]
    outputs = [state, chat, prompt, output]

    prompt.submit(generate, inputs=inputs, outputs=outputs)
    button_generate.click(generate, inputs=inputs, outputs=outputs)
    button_reset.click(reset, inputs=[state], outputs=[state, chat, output])

    examples = gr.Examples(
        inputs=[context, prompt, model, do_sample, top_k, top_p, temperature],
        examples=[
            [
                "Human talks to a powerful AI that follows the Human's instructions. "
                "AI is a smart, talkative, friendly, honest, helpful, harmless assistant to Human. "
                "AI has instant access to an online encyclopedia containing all the facts about the world "
                "and answers any question in detail. AI never says common misconceptions, "
                "outdated information, lies, fiction, myths, jokes, or memes.</s>\n"
                "AI: Hi! How can I help you?</s>\n",
                "Could you remind me please who was Neil Armstrong?",
                "stabilityai/StableBeluga2",
                True,
                0,
                0.9,
                0.75,
            ],
            [
                "Human mluví s mocnou, inteligentní a vševědoucí AI, která plní instrukce od Human. "
                "AI je výřečná, přátelská, pozitivní a poskytuje detailní odpovědi na jakoukoliv otázku.</s>\n"
                "Human: Ahoj!</s>\n"
                "AI: Ahoj! Jak ti mohu pomoci?",
                "Můžeš mi prosím připomenout, kdo byl Neil Armstrong?",
                "stabilityai/StableBeluga2",
                True,
                0,
                0.9,
                0.75,
            ],
         ],
    )