I use this link to learn object detection on windows 10.
I prepared 400 pictures and divided them into two classes(stones and cars).
Then I used this command to train:
cd E:\test\models-master\research\object_detection
python model_main.py --pipeline_config_path=training/ssd_mobilenet_v1_coco.config --model_dir=training/ --num_train_steps=10000
In object_detection/model_main.py I see a parameter called checkpoint_dir.
But I don't know how to use checkpoint_dir.If my model is trained to more than 6000 steps, the training folder looks like the following picture:
And then I stop training the model.When I want to continue training, how to set checkpoint_dir?
I use this command:
python model_main.py --pipeline_config_path=training/ssd_mobilenet_v1_coco.config -- model_dir=training/ --checkpoint_dir=training/ --num_train_steps=20000 --alsologtostderr
When I added --checkpoint_dir=training/, the model didn't continue training. Why?How to use --checkpoint_dir?
I download ssd_mobilenet_v1_coco_2018_01_28.tar.gz from the detection_model_zoo.
Then I unzipped ssd_mobilenet_v1_coco_2018_01_28.tar.gz to folder object_detection/ssd_mobilenet_v1_coco_2018_01_28.
The object_detection/ssd_mobilenet_v1_coco_2018_01_28 folder has files like this：
So how to use fine_tune_checkpoint in training/ssd_mobilenet_v1_coco.config?
The content in training/ssd_mobilenet_v1_coco.config looks like this:
# SSD with Mobilenet v1 configuration for MSCOCO Dataset.
# Users should configure the fine_tune_checkpoint field in the train config as
# well as the label_map_path and input_path fields in the train_input_reader and
# eval_input_reader. Search for "PATH_TO_BE_CONFIGURED" to find the fields that
# should be configured.
model {
ssd {
num_classes: 2
box_coder {
faster_rcnn_box_coder {
y_scale: 10.0
x_scale: 10.0
height_scale: 5.0
width_scale: 5.0
}
}
matcher {
argmax_matcher {
matched_threshold: 0.5
unmatched_threshold: 0.5
ignore_thresholds: false
negatives_lower_than_unmatched: true
force_match_for_each_row: true
}
}
similarity_calculator {
iou_similarity {
}
}
anchor_generator {
ssd_anchor_generator {
num_layers: 6
min_scale: 0.2
max_scale: 0.95
aspect_ratios: 1.0
aspect_ratios: 2.0
aspect_ratios: 0.5
aspect_ratios: 3.0
aspect_ratios: 0.3333
}
}
image_resizer {
fixed_shape_resizer {
height: 300
width: 300
}
}
box_predictor {
convolutional_box_predictor {
min_depth: 0
max_depth: 0
num_layers_before_predictor: 0
use_dropout: false
dropout_keep_probability: 0.8
kernel_size: 1
box_code_size: 4
apply_sigmoid_to_scores: false
conv_hyperparams {
activation: RELU_6,
regularizer {
l2_regularizer {
weight: 0.00004
}
}
initializer {
truncated_normal_initializer {
stddev: 0.03
mean: 0.0
}
}
batch_norm {
train: true,
scale: true,
center: true,
decay: 0.9997,
epsilon: 0.001,
}
}
}
}
feature_extractor {
type: 'ssd_mobilenet_v1'
min_depth: 16
depth_multiplier: 1.0
conv_hyperparams {
activation: RELU_6,
regularizer {
l2_regularizer {
weight: 0.00004
}
}
initializer {
truncated_normal_initializer {
stddev: 0.03
mean: 0.0
}
}
batch_norm {
train: true,
scale: true,
center: true,
decay: 0.9997,
epsilon: 0.001,
}
}
}
loss {
classification_loss {
weighted_sigmoid {
}
}
localization_loss {
weighted_smooth_l1 {
}
}
hard_example_miner {
num_hard_examples: 3000
iou_threshold: 0.99
loss_type: CLASSIFICATION
max_negatives_per_positive: 3
min_negatives_per_image: 0
}
classification_weight: 1.0
localization_weight: 1.0
}
normalize_loss_by_num_matches: true
post_processing {
batch_non_max_suppression {
score_threshold: 1e-8
iou_threshold: 0.6
max_detections_per_class: 100
max_total_detections: 100
}
score_converter: SIGMOID
}
}
}
train_config: {
batch_size: 10
optimizer {
rms_prop_optimizer: {
learning_rate: {
exponential_decay_learning_rate {
initial_learning_rate: 0.004
decay_steps: 800720
decay_factor: 0.95
}
}
momentum_optimizer_value: 0.9
decay: 0.9
epsilon: 1.0
}
}
fine_tune_checkpoint: 'ssd_mobilenet_v1_coco_2018_01_28/model.ckpt'
from_detection_checkpoint: true
# Note: The below line limits the training process to 200K steps, which we
# empirically found to be sufficient enough to train the pets dataset. This
# effectively bypasses the learning rate schedule (the learning rate will
# never decay). Remove the below line to train indefinitely.
num_steps: 1000
data_augmentation_options {
random_horizontal_flip {
}
}
data_augmentation_options {
ssd_random_crop {
}
}
}
train_input_reader: {
tf_record_input_reader {
input_path:'data/train.record'
}
label_map_path:'data/side_vehicle.pbtxt'
}
eval_config: {
num_examples: 8000
# Note: The below line limits the evaluation process to 10 evaluations.
# Remove the below line to evaluate indefinitely.
max_evals: 10
}
eval_input_reader: {
tf_record_input_reader {
input_path: 'data/test.record'
}
label_map_path: 'data/side_vehicle.pbtxt'
shuffle: false
num_readers: 1
}
Are these two lines right?
fine_tune_checkpoint: 'ssd_mobilenet_v1_coco_2018_01_28/model.ckpt'
from_detection_checkpoint: true
What is the difference between checkpoint_dir and fine_tune_checkpoint in tensorflow object detection?
