292 lines
7.7 KiB
Python
Executable File
292 lines
7.7 KiB
Python
Executable File
#!/usr/bin/env python3
|
|
#
|
|
# train.py
|
|
#
|
|
# SPDX-License-Identifier: BSD-2-Clause
|
|
#
|
|
# Copyright (c) 2023, Jeff Moe
|
|
# All rights reserved.
|
|
#
|
|
# Redistribution and use in source and binary forms, with or without
|
|
# modification, are permitted provided that the following conditions are met:
|
|
#
|
|
# 1. Redistributions of source code must retain the above copyright notice, this
|
|
# list of conditions and the following disclaimer.
|
|
#
|
|
# 2. Redistributions in binary form must reproduce the above copyright notice,
|
|
# this list of conditions and the following disclaimer in the documentation
|
|
# and/or other materials provided with the distribution.
|
|
#
|
|
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
|
|
# AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
|
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
|
|
# DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE
|
|
# FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
|
|
# DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
|
|
# SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
|
|
# CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
|
|
# OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
|
|
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
|
|
#
|
|
|
|
import argparse
|
|
import subprocess
|
|
|
|
# PyTorch version used in paper ?
|
|
# Python version used in paper ?
|
|
# Set hyperparameters based on values in Table 3 of
|
|
# Pl@ntNet-300K paper.
|
|
#
|
|
# Use defaults from git repo example.
|
|
# https://github.com/plantnet/PlantNet-300K
|
|
BATCH_SIZE = "32"
|
|
MU = "0.0001"
|
|
KONE = "1"
|
|
KTWO = "3"
|
|
KTHREE = "5"
|
|
KFOUR = "10"
|
|
SEED = "4"
|
|
IMAGE_SIZE = "256"
|
|
CROP_SIZE = "224"
|
|
|
|
# Root path to images test train val
|
|
ROOT_DIR = "/srv/ml/plantnet/files/plantnet_300K/images"
|
|
# Use GPU
|
|
USE_GPU = "1"
|
|
# Use all CPUs available on system. XXX get nproc
|
|
NUM_WORKERS = "12"
|
|
|
|
# Parse command line options
|
|
parser = argparse.ArgumentParser(
|
|
prog="train.py",
|
|
description="Train PlantNet-300K models using default parameters.",
|
|
epilog="Example: ./train.py alexnet",
|
|
)
|
|
|
|
parser.add_argument(
|
|
"model",
|
|
help="Model name",
|
|
type=str,
|
|
choices=[
|
|
"alexnet",
|
|
"densenet121",
|
|
"densenet161",
|
|
"densenet169",
|
|
"densenet201",
|
|
"efficientnet_b0",
|
|
"efficientnet_b1",
|
|
"efficientnet_b2",
|
|
"efficientnet_b3",
|
|
"efficientnet_b4",
|
|
"inception_resnet_v2",
|
|
"inception_v3",
|
|
"inception_v4",
|
|
"mobilenet_v2",
|
|
"mobilenet_v3_large",
|
|
"mobilenet_v3_small",
|
|
"resnet18",
|
|
"resnet34",
|
|
"resnet50",
|
|
"resnet101",
|
|
"resnet152",
|
|
"shufflenet_v2_x1_0",
|
|
"squeezenet1_0",
|
|
"vgg11",
|
|
"vit_b_16",
|
|
"wide_resnet50_2",
|
|
"wide_resnet101_2",
|
|
],
|
|
)
|
|
args = parser.parse_args()
|
|
MODEL_NAME = args.model
|
|
|
|
# Initial Learning Rate
|
|
LR = "N.NNNN"
|
|
# Number of Epochs
|
|
N_EPOCHS = "NN"
|
|
# First Decay
|
|
FIRST_DECAY = "1N"
|
|
# Secon Decay
|
|
SECOND_DECAY = "2N"
|
|
# Set LR, epochs, decay hyperparameters based on model.
|
|
match MODEL_NAME:
|
|
case "alexnet":
|
|
LR = "0.001"
|
|
N_EPOCHS = "30"
|
|
FIRST_DECAY = "20"
|
|
SECOND_DECAY = "25"
|
|
case "densenet121":
|
|
LR = "0.01"
|
|
N_EPOCHS = "30"
|
|
FIRST_DECAY = "20"
|
|
SECOND_DECAY = "25"
|
|
case "densenet161":
|
|
LR = "0.01"
|
|
N_EPOCHS = "30"
|
|
FIRST_DECAY = "20"
|
|
SECOND_DECAY = "25"
|
|
case "densenet169":
|
|
LR = "0.01"
|
|
N_EPOCHS = "30"
|
|
FIRST_DECAY = "20"
|
|
SECOND_DECAY = "25"
|
|
case "densenet201":
|
|
LR = "0.01"
|
|
N_EPOCHS = "30"
|
|
FIRST_DECAY = "20"
|
|
SECOND_DECAY = "25"
|
|
case "efficientnet_b0":
|
|
LR = "0.01"
|
|
N_EPOCHS = "20"
|
|
FIRST_DECAY = "10"
|
|
SECOND_DECAY = "15"
|
|
case "efficientnet_b1":
|
|
LR = "0.01"
|
|
N_EPOCHS = "20"
|
|
FIRST_DECAY = "10"
|
|
SECOND_DECAY = "15"
|
|
case "efficientnet_b2":
|
|
LR = "0.01"
|
|
N_EPOCHS = "20"
|
|
FIRST_DECAY = "10"
|
|
SECOND_DECAY = "15"
|
|
case "efficientnet_b3":
|
|
LR = "0.01"
|
|
N_EPOCHS = "20"
|
|
FIRST_DECAY = "10"
|
|
SECOND_DECAY = "15"
|
|
case "efficientnet_b4":
|
|
LR = "0.01"
|
|
N_EPOCHS = "20"
|
|
FIRST_DECAY = "10"
|
|
SECOND_DECAY = "15"
|
|
case "inception_resnet_v2":
|
|
LR = "0.01"
|
|
N_EPOCHS = "30"
|
|
FIRST_DECAY = "20"
|
|
SECOND_DECAY = "25"
|
|
case "inception_v3":
|
|
LR = "0.01"
|
|
N_EPOCHS = "30"
|
|
FIRST_DECAY = "20"
|
|
SECOND_DECAY = "25"
|
|
case "inception_v4":
|
|
LR = "0.01"
|
|
N_EPOCHS = "30"
|
|
FIRST_DECAY = "20"
|
|
SECOND_DECAY = "25"
|
|
case "mobilenet_v2":
|
|
LR = "0.01"
|
|
N_EPOCHS = "30"
|
|
FIRST_DECAY = "20"
|
|
SECOND_DECAY = "25"
|
|
case "mobilenet_v3_large":
|
|
LR = "0.01"
|
|
N_EPOCHS = "30"
|
|
FIRST_DECAY = "20"
|
|
SECOND_DECAY = "25"
|
|
case "mobilenet_v3_small":
|
|
LR = "0.001"
|
|
N_EPOCHS = "30"
|
|
FIRST_DECAY = "20"
|
|
SECOND_DECAY = "25"
|
|
case "resnet18":
|
|
LR = "0.01"
|
|
N_EPOCHS = "30"
|
|
FIRST_DECAY = "20"
|
|
SECOND_DECAY = "25"
|
|
case "resnet34":
|
|
LR = "0.01"
|
|
N_EPOCHS = "30"
|
|
FIRST_DECAY = "20"
|
|
SECOND_DECAY = "25"
|
|
case "resnet50":
|
|
LR = "0.01"
|
|
N_EPOCHS = "30"
|
|
FIRST_DECAY = "20"
|
|
SECOND_DECAY = "25"
|
|
case "resnet101":
|
|
LR = "0.01"
|
|
N_EPOCHS = "30"
|
|
FIRST_DECAY = "20"
|
|
SECOND_DECAY = "25"
|
|
case "resnet152":
|
|
LR = "0.01"
|
|
N_EPOCHS = "30"
|
|
FIRST_DECAY = "20"
|
|
SECOND_DECAY = "25"
|
|
case "shufflenet_v2_x1_0":
|
|
LR = "0.01"
|
|
N_EPOCHS = "30"
|
|
FIRST_DECAY = "20"
|
|
SECOND_DECAY = "25"
|
|
case "squeezenet1_0":
|
|
LR = "0.001"
|
|
N_EPOCHS = "30"
|
|
FIRST_DECAY = "20"
|
|
SECOND_DECAY = "25"
|
|
case "vgg11":
|
|
LR = "0.001"
|
|
N_EPOCHS = "30"
|
|
FIRST_DECAY = "20"
|
|
SECOND_DECAY = "25"
|
|
case "vit_b_16":
|
|
LR = "0.0005"
|
|
N_EPOCHS = "20"
|
|
FIRST_DECAY = "15"
|
|
SECOND_DECAY = ""
|
|
case "wide_resnet50_2":
|
|
LR = "0.01"
|
|
N_EPOCHS = "30"
|
|
FIRST_DECAY = "20"
|
|
SECOND_DECAY = "25"
|
|
case "wide_resnet101_2":
|
|
LR = "0.01"
|
|
N_EPOCHS = "30"
|
|
FIRST_DECAY = "20"
|
|
SECOND_DECAY = "25"
|
|
|
|
# Print command to run
|
|
print(
|
|
"python",
|
|
"main.py",
|
|
"--model=" + MODEL_NAME,
|
|
"--lr=" + LR,
|
|
"--n_epochs=" + N_EPOCHS,
|
|
"--epoch_decay=" + FIRST_DECAY + SECOND_DECAY,
|
|
"--k=" + str(KONE) + str(KTWO) + str(KTHREE) + str(KFOUR),
|
|
"--batch_size=" + BATCH_SIZE,
|
|
"--mu=" + MU,
|
|
"--pretrained",
|
|
"--seed=" + SEED,
|
|
"--image_size=" + IMAGE_SIZE,
|
|
"--crop_size=" + CROP_SIZE,
|
|
"--root=" + ROOT_DIR,
|
|
"--use_gpu=" + USE_GPU,
|
|
"--num_workers=" + NUM_WORKERS,
|
|
"--save_name_xp=" + MODEL_NAME,
|
|
)
|
|
|
|
# Run model with corresponding hyperparameters
|
|
subprocess.run(
|
|
[
|
|
"python",
|
|
"main.py",
|
|
"--model=" + MODEL_NAME,
|
|
"--lr=" + LR,
|
|
"--n_epochs=" + N_EPOCHS,
|
|
"--epoch_decay=" + FIRST_DECAY + SECOND_DECAY,
|
|
"--k=" + str(KONE) + str(KTWO) + str(KTHREE) + str(KFOUR),
|
|
"--batch_size=" + BATCH_SIZE,
|
|
"--mu=" + MU,
|
|
"--pretrained",
|
|
"--seed=" + SEED,
|
|
"--image_size=" + IMAGE_SIZE,
|
|
"--crop_size=" + CROP_SIZE,
|
|
"--root=" + ROOT_DIR,
|
|
"--use_gpu=" + USE_GPU,
|
|
"--num_workers=" + NUM_WORKERS,
|
|
"--save_name_xp=" + MODEL_NAME,
|
|
]
|
|
)
|