mikeee's picture
Duplicate from mikeee/llama-2-70b-guanaco-qlora-ggml
64dc528
raw
history blame
11.7 kB
"""Run codes."""
# pylint: disable=line-too-long, broad-exception-caught, invalid-name, missing-function-docstring, too-many-instance-attributes, missing-class-docstring
# ruff: noqa: E501
import os
import platform
import random
import time
from dataclasses import asdict, dataclass, field
from pathlib import Path
from textwrap import dedent
# from types import SimpleNamespace
import gradio as gr
import psutil
from about_time import about_time
from ctransformers import AutoModelForCausalLM
from dl_hf_model import dl_hf_model
from loguru import logger
from examples_list import examples_list
url = "https://huggingface.co./TheBloke/llama-2-13B-Guanaco-QLoRA-GGML/blob/main/llama-2-13b-guanaco-qlora.ggmlv3.q4_K_S.bin" # 8.14G
LLM = None
if "forindo" in platform.node(): # deploy 70b model locally
# url = "https://huggingface.co./TheBloke/llama-2-70b-Guanaco-QLoRA-GGML/blob/main/llama-2-70b-guanaco-qlora.ggmlv3.q3_K_S.bin" # 29.7G
# model_loc = "/home/mu2018/github/langchain-llama-2-70b-guanaco-qlora-ggml/models/llama-2-70b-guanaco-qlora.ggmlv3.q3_K_S.bin"
_ = """
url = "https://huggingface.co./TheBloke/StableBeluga2-70B-GGML/blob/main/stablebeluga2-70b.ggmlv3.q3_K_S.bin"
try:
model_loc, file_size = dl_hf_model(url)
logger.info(f"done load llm {model_loc=} {file_size=}G")
except Exception as exc_:
logger.error(exc_)
raise SystemExit(1) from exc_
# """
model_loc = "models/stablebeluga2-70b.ggmlv3.q3_K_S.bin"
assert Path(model_loc).exists(), f"Make sure {model_loc=} exists."
else:
try:
logger.debug(f" dl {url}")
model_loc, file_size = dl_hf_model(url)
logger.info(f"done load llm {model_loc=} {file_size=}G")
except Exception as exc_:
logger.error(exc_)
raise SystemExit(1) from exc_
# raise SystemExit(0)
# Prompt template: Guanaco
# {past_history}
prompt_template = """You are a helpful assistant. Let's think step by step.
### Human:
{question}
### Assistant:"""
human_prefix = "### Human"
ai_prefix = "### Assistant"
stop_list = [f"{human_prefix}:"]
if "beluga" in model_loc.lower():
prompt_template = dedent(
"""
### System:
You are Stable Beluga, an AI that follows instructions extremely well. Help as much as you can.
Let's think step by step.
### User: {question}
### Assistant:
"""
).lstrip()
human_prefix = "### User"
ai_prefix = "### Assistant"
stop_list = [f"{human_prefix}:"]
_ = psutil.cpu_count(logical=False) - 1
cpu_count: int = int(_) if _ else 1
logger.debug(f"{cpu_count=}")
logger.debug(f"{model_loc=}")
LLM = AutoModelForCausalLM.from_pretrained(
model_loc,
model_type="llama",
threads=cpu_count,
)
os.environ["TZ"] = "Asia/Shanghai"
try:
time.tzset() # type: ignore # pylint: disable=no-member
except Exception:
# Windows
logger.warning("Windows, cant run time.tzset()")
@dataclass
class GenerationConfig:
temperature: float = 0.7
top_k: int = 50
top_p: float = 0.9
repetition_penalty: float = 1.0
max_new_tokens: int = 512
seed: int = 42
reset: bool = False
stream: bool = True
threads: int = cpu_count
stop: list[str] = field(default_factory=lambda: stop_list)
def generate(
question: str,
llm=LLM,
config: GenerationConfig = GenerationConfig(),
):
"""Run model inference, will return a Generator if streaming is true."""
# _ = prompt_template.format(question=question)
# print(_)
prompt = prompt_template.format(question=question)
return llm(
prompt,
**asdict(config),
)
logger.debug(f"{asdict(GenerationConfig())=}")
def user(user_message, history):
# return user_message, history + [[user_message, None]]
if history is None:
history = []
history.append([user_message, None])
return user_message, history # keep user_message
def user1(user_message, history):
# return user_message, history + [[user_message, None]]
if history is None:
history = []
history.append([user_message, None])
return "", history # clear user_message
def bot_(history):
user_message = history[-1][0]
resp = random.choice(["How are you?", "I love you", "I'm very hungry"])
bot_message = user_message + ": " + resp
history[-1][1] = ""
for character in bot_message:
history[-1][1] += character
time.sleep(0.02)
yield history
history[-1][1] = resp
yield history
def bot(history):
user_message = ""
try:
user_message = history[-1][0]
except Exception as exc:
logger.error(exc)
response = []
logger.debug(f"{user_message=}")
with about_time() as atime: # type: ignore
flag = 1
prefix = ""
then = time.time()
logger.debug("about to generate")
config = GenerationConfig(reset=True)
for elm in generate(user_message, config=config):
if flag == 1:
logger.debug("in the loop")
prefix = f"({time.time() - then:.2f}s) "
flag = 0
print(prefix, end="", flush=True)
logger.debug(f"{prefix=}")
print(elm, end="", flush=True)
# logger.debug(f"{elm}")
response.append(elm)
history[-1][1] = prefix + "".join(response)
yield history
_ = (
f"(time elapsed: {atime.duration_human}, " # type: ignore
f"{atime.duration/len(''.join(response)):.2f}s/char)" # type: ignore
)
history[-1][1] = "".join(response) + f"\n{_}"
yield history
def predict_api(prompt):
logger.debug(f"{prompt=}")
try:
# user_prompt = prompt
config = GenerationConfig(
temperature=0.2,
top_k=10,
top_p=0.9,
repetition_penalty=1.0,
max_new_tokens=512, # adjust as needed
seed=42,
reset=True, # reset history (cache)
stream=False,
# threads=cpu_count,
# stop=prompt_prefix[1:2],
)
response = generate(
prompt,
config=config,
)
logger.debug(f"api: {response=}")
except Exception as exc:
logger.error(exc)
response = f"{exc=}"
# bot = {"inputs": [response]}
# bot = [(prompt, response)]
return response
css = """
.importantButton {
background: linear-gradient(45deg, #7e0570,#5d1c99, #6e00ff) !important;
border: none !important;
}
.importantButton:hover {
background: linear-gradient(45deg, #ff00e0,#8500ff, #6e00ff) !important;
border: none !important;
}
.disclaimer {font-variant-caps: all-small-caps; font-size: xx-small;}
.xsmall {font-size: x-small;}
"""
logger.info("start block")
with gr.Blocks(
title=f"{Path(model_loc).name}",
# theme=gr.themes.Soft(text_size="sm", spacing_size="sm"),
theme=gr.themes.Glass(text_size="sm", spacing_size="sm"),
css=css,
) as block:
# buff_var = gr.State("")
with gr.Accordion("🎈 Info", open=False):
gr.Markdown(
f"""<h5><center>{Path(model_loc).name}</center></h4>
Most examples are meant for another model.
You probably should try to test
some related prompts.""",
elem_classes="xsmall",
)
# chatbot = gr.Chatbot().style(height=700) # 500
chatbot = gr.Chatbot(height=500)
# buff = gr.Textbox(show_label=False, visible=True)
with gr.Row():
with gr.Column(scale=5):
msg = gr.Textbox(
label="Chat Message Box",
placeholder="Ask me anything (press Shift+Enter or click Submit to send)",
show_label=False,
# container=False,
lines=6,
max_lines=30,
show_copy_button=True,
# ).style(container=False)
)
with gr.Column(scale=1, min_width=50):
with gr.Row():
submit = gr.Button("Submit", elem_classes="xsmall")
stop = gr.Button("Stop", visible=True)
clear = gr.Button("Clear History", visible=True)
with gr.Row(visible=False):
with gr.Accordion("Advanced Options:", open=False):
with gr.Row():
with gr.Column(scale=2):
system = gr.Textbox(
label="System Prompt",
value=prompt_template,
show_label=False,
container=False,
# ).style(container=False)
)
with gr.Column():
with gr.Row():
change = gr.Button("Change System Prompt")
reset = gr.Button("Reset System Prompt")
with gr.Accordion("Example Inputs", open=True):
examples = gr.Examples(
examples=examples_list,
inputs=[msg],
examples_per_page=40,
)
# with gr.Row():
with gr.Accordion("Disclaimer", open=False):
_ = Path(model_loc).name
gr.Markdown(
f"Disclaimer: {_} can produce factually incorrect output, and should not be relied on to produce "
"factually accurate information. {_} 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"],
)
msg_submit_event = msg.submit(
# fn=conversation.user_turn,
fn=user,
inputs=[msg, chatbot],
outputs=[msg, chatbot],
queue=True,
show_progress="full",
# api_name=None,
).then(bot, chatbot, chatbot, queue=True)
submit_click_event = submit.click(
# fn=lambda x, y: ("",) + user(x, y)[1:], # clear msg
fn=user1, # clear msg
inputs=[msg, chatbot],
outputs=[msg, chatbot],
queue=True,
# queue=False,
show_progress="full",
# api_name=None,
).then(bot, chatbot, chatbot, queue=True)
stop.click(
fn=None,
inputs=None,
outputs=None,
cancels=[msg_submit_event, submit_click_event],
queue=False,
)
clear.click(lambda: None, None, chatbot, queue=False)
with gr.Accordion("For Chat/Translation API", open=False, visible=False):
input_text = gr.Text()
api_btn = gr.Button("Go", variant="primary")
out_text = gr.Text()
api_btn.click(
predict_api,
input_text,
out_text,
api_name="api",
)
# block.load(update_buff, [], buff, every=1)
# block.load(update_buff, [buff_var], [buff_var, buff], every=1)
# concurrency_count=5, max_size=20
# max_size=36, concurrency_count=14
# CPU cpu_count=2 16G, model 7G
# CPU UPGRADE cpu_count=8 32G, model 7G
# does not work
_ = """
# _ = int(psutil.virtual_memory().total / 10**9 // file_size - 1)
# concurrency_count = max(_, 1)
if psutil.cpu_count(logical=False) >= 8:
# concurrency_count = max(int(32 / file_size) - 1, 1)
else:
# concurrency_count = max(int(16 / file_size) - 1, 1)
# """
# default concurrency_count = 1
# block.queue(concurrency_count=concurrency_count, max_size=5).launch(debug=True)
server_port = 7860
if "forindo" in platform.node():
server_port = 7861
block.queue(max_size=5).launch(
debug=True, server_name="0.0.0.0", server_port=server_port
)
# block.queue(max_size=5).launch(debug=True, server_name="0.0.0.0")