Route generate

trainer/src/lora_inference/experiments.py:None–None

stub.function(
    gpu=modal.gpu.A10G(count=2),
    secret=modal.Secret.from_name("openpipe"),
    allow_concurrent_inputs=12,
    keep_warm=1,
    timeout=1 * 60 * 60,
    volumes={"/models": volume}

import modal

from .api import (
    Input,