tabby/website/docs/installation/modal/app.py

84 lines
1.8 KiB
Python

"""Usage:
modal serve app.py
"""
from modal import Image, Stub, asgi_app, gpu
IMAGE_NAME = "tabbyml/tabby:0.6.0"
MODEL_ID = "TabbyML/StarCoder-1B"
GPU_CONFIG = gpu.T4()
def download_model():
import subprocess
subprocess.run(
[
"/opt/tabby/bin/tabby",
"download",
"--model",
MODEL_ID,
]
)
image = (
Image.from_registry(
IMAGE_NAME,
add_python="3.11",
)
.dockerfile_commands("ENTRYPOINT []")
.run_function(download_model)
.pip_install("asgi-proxy-lib")
)
stub = Stub("tabby-server-" + MODEL_ID.split("/")[-1], image=image)
@stub.function(
gpu=GPU_CONFIG,
allow_concurrent_inputs=10,
container_idle_timeout=120,
timeout=360,
)
@asgi_app()
def app():
import socket
import subprocess
import time
from asgi_proxy import asgi_proxy
launcher = subprocess.Popen(
[
"/opt/tabby/bin/tabby",
"serve",
"--model",
MODEL_ID,
"--port",
"8000",
"--device",
"cuda",
"--parallelism",
"4",
]
)
# Poll until webserver at 127.0.0.1:8000 accepts connections before running inputs.
def tabby_ready():
try:
socket.create_connection(("127.0.0.1", 8000), timeout=1).close()
return True
except (socket.timeout, ConnectionRefusedError):
# Check if launcher webserving process has exited.
# If so, a connection can never be made.
retcode = launcher.poll()
if retcode is not None:
raise RuntimeError(f"launcher exited unexpectedly with code {retcode}")
return False
while not tabby_ready():
time.sleep(1.0)
print("Tabby server ready!")
return asgi_proxy("http://localhost:8000")