# Copyright 2023 MosaicML spaces authors # SPDX-License-Identifier: Apache-2.0 # and # the https://huggingface.co/spaces/HuggingFaceH4/databricks-dolly authors import datetime import os from threading import Event, Thread from uuid import uuid4 import gradio as gr import requests import torch from transformers import StoppingCriteria, StoppingCriteriaList, TextIteratorStreamer from quick_pipeline import InstructionTextGenerationPipeline as pipeline # Configuration HF_TOKEN = os.getenv("HF_TOKEN", None) examples = [ # to do: add coupled hparams so e.g. poem has higher temp "Write a travel blog about a 3-day trip to Thailand.", "Write a short story about a robot that has a nice day.", "Convert the following to a single line of JSON:\n\n```name: John\nage: 30\naddress:\n street:123 Main St.\n city: San Francisco\n state: CA\n zip: 94101\n```", "Write a quick email to congratulate MosaicML about the launch of their inference offering.", "Explain how a candle works to a 6 year old in a few sentences.", "What are some of the most common misconceptions about birds?", ] # Initialize the model and tokenizer generate = pipeline( "mosaicml/mpt-7b-instruct", torch_dtype=torch.bfloat16, trust_remote_code=True, use_auth_token=HF_TOKEN, ) stop_token_ids = generate.tokenizer.convert_tokens_to_ids(["<|endoftext|>"]) # Define a custom stopping criteria class StopOnTokens(StoppingCriteria): def __call__(self, input_ids: torch.LongTensor, scores: torch.FloatTensor, **kwargs) -> bool: for stop_id in stop_token_ids: if input_ids[0][-1] == stop_id: return True return False def log_conversation(session_id, instruction, response, generate_kwargs): logging_url = os.getenv("LOGGING_URL", None) if logging_url is None: return timestamp = datetime.datetime.now().strftime("%Y-%m-%dT%H:%M:%S") data = { "session_id": session_id, "timestamp": timestamp, "instruction": instruction, "response": response, "generate_kwargs": generate_kwargs, } try: requests.post(logging_url, json=data) except requests.exceptions.RequestException as e: print(f"Error logging conversation: {e}") def process_stream(instruction, temperature, top_p, top_k, max_new_tokens, session_id): # Tokenize the input input_ids = generate.tokenizer( generate.format_instruction(instruction), return_tensors="pt" ).input_ids input_ids = input_ids.to(generate.model.device) # Initialize the streamer and stopping criteria streamer = TextIteratorStreamer( generate.tokenizer, timeout=10.0, skip_prompt=True, skip_special_tokens=True ) stop = StopOnTokens() if temperature < 0.1: temperature = 0.0 do_sample = False else: do_sample = True gkw = { **generate.generate_kwargs, **{ "input_ids": input_ids, "max_new_tokens": max_new_tokens, "temperature": temperature, "do_sample": do_sample, "top_p": top_p, "top_k": top_k, "streamer": streamer, "stopping_criteria": StoppingCriteriaList([stop]), }, } response = "" stream_complete = Event() def generate_and_signal_complete(): generate.model.generate(**gkw) stream_complete.set() def log_after_stream_complete(): stream_complete.wait() log_conversation( session_id, instruction, response, { "top_k": top_k, "top_p": top_p, "temperature": temperature, }, ) t1 = Thread(target=generate_and_signal_complete) t1.start() t2 = Thread(target=log_after_stream_complete) t2.start() for new_text in streamer: response += new_text yield response with gr.Blocks( theme=gr.themes.Soft(), css=".disclaimer {font-variant-caps: all-small-caps;}", ) as demo: session_id = gr.State(lambda: str(uuid4())) gr.Markdown( """

MosaicML MPT-7B-Instruct

This demo is of [MPT-7B-Instruct](https://huggingface.co/mosaicml/mpt-7b-instruct). It is based on [MPT-7B](https://huggingface.co/mosaicml/mpt-7b) fine-tuned with approximately [60,000 instruction demonstrations](https://huggingface.co/datasets/sam-mosaic/dolly_hhrlhf) If you're interested in [training](https://www.mosaicml.com/training) and [deploying](https://www.mosaicml.com/inference) your own MPT or LLMs, [sign up](https://forms.mosaicml.com/demo?utm_source=huggingface&utm_medium=referral&utm_campaign=mpt-7b) for MosaicML platform. This is running on a smaller, shared GPU, so it may take a few seconds to respond. If you want to run it on your own GPU, you can [download the model from HuggingFace](https://huggingface.co/mosaicml/mpt-7b-instruct) and run it locally. Or [Duplicate the Space](https://huggingface.co/spaces/mosaicml/mpt-7b-instruct?duplicate=true) to skip the queue and run in a private space.""" ) with gr.Row(): with gr.Column(): with gr.Row(): instruction = gr.Textbox( placeholder="Enter your question here", label="Question/Instruction", elem_id="q-input", ) with gr.Accordion("Advanced Options:", open=False): with gr.Row(): with gr.Column(): with gr.Row(): temperature = gr.Slider( label="Temperature", value=0.1, minimum=0.0, maximum=1.0, step=0.1, interactive=True, info="Higher values produce more diverse outputs", ) with gr.Column(): with gr.Row(): top_p = gr.Slider( label="Top-p (nucleus sampling)", value=1.0, minimum=0.0, maximum=1, step=0.01, interactive=True, info=( "Sample from the smallest possible set of tokens whose cumulative probability " "exceeds top_p. Set to 1 to disable and sample from all tokens." ), ) with gr.Column(): with gr.Row(): top_k = gr.Slider( label="Top-k", value=0, minimum=0.0, maximum=200, step=1, interactive=True, info="Sample from a shortlist of top-k tokens — 0 to disable and sample from all tokens.", ) with gr.Column(): with gr.Row(): max_new_tokens = gr.Slider( label="Maximum new tokens", value=256, minimum=0, maximum=1664, step=5, interactive=True, info="The maximum number of new tokens to generate", ) with gr.Row(): submit = gr.Button("Submit") with gr.Row(): with gr.Box(): gr.Markdown("**MPT-7B-Instruct**") output_7b = gr.Markdown() with gr.Row(): gr.Examples( examples=examples, inputs=[instruction], cache_examples=False, fn=process_stream, outputs=output_7b, ) with gr.Row(): gr.Markdown( "Disclaimer: MPT-7B can produce factually incorrect output, and should not be relied on to produce " "factually accurate information. MPT-7B was trained on various public datasets; while great efforts " "have been taken to clean the pretraining data, it is possible that this model could generate lewd, " "biased, or otherwise offensive outputs.", elem_classes=["disclaimer"], ) with gr.Row(): gr.Markdown( "[Privacy policy](https://gist.github.com/samhavens/c29c68cdcd420a9aa0202d0839876dac)", elem_classes=["disclaimer"], ) submit.click( process_stream, inputs=[instruction, temperature, top_p, top_k, max_new_tokens, session_id], outputs=output_7b, ) instruction.submit( process_stream, inputs=[instruction, temperature, top_p, top_k, max_new_tokens, session_id], outputs=output_7b, ) demo.queue(max_size=32, concurrency_count=4).launch(debug=True)