https://github.com/tensorflow/models/tree/master/research/object_detection
tensorflow中的models模塊中research中的object detection模塊
object detection模塊看似和slim框架平級但是其實是在object detection模塊有使用slim框架的部分
所以object detection模塊不是獨立運行的,我們認爲object detection模塊基於slim框架的。
數據集:
數據的介紹
Image:存的是圖片
Xml:圖片的信息和標定框的信息
Boundiningbox在這裏使用的是像素
Labels_item.txt
模型的類別與數字對應的關係
數據的準備
1設定環境變量
我們需要設定object-detection和tf-slim的環境變量才能使用object-detection和tf-slim。
export PYTHONPATH=$PYTHONPATH:./models/research:./models/research/slim
2預編譯文件
這些文件就是我們運行程序之前需要預編譯的(proto後綴)
在research目錄下 執行下面程序 預編譯文件
/models/research$ protoc object_detection/protos/*.proto --python_out=.
轉換數據
python3 quiz-object-detection/create_data.py\
--data_dir ./quiz-w8-data\
--label_map_path ./quiz-w8-data/labels_items.txt\
--output_dir ./
最後生成tf-record文件
訓練準備
1下載檢測的預訓練模型
下載好的預訓練模型
解壓下面就是我們的預訓練模型
2把這些文件放在data文件夾中
3修改我們要使用模型的對應的配置文件
找到與我們預訓練模型先對應的配置文件
修改成如下
# SSD with Mobilenet v1, configured for Oxford-IIIT Pets 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 "/home/dl/ai100/quiz-w8-data" to find the fields that
# should be configured.
model {
ssd {
num_classes: 5
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 {
anchorwise_output: true
}
}
localization_loss {
weighted_smooth_l1 {
anchorwise_output: true
}
}
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: 24
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: "/data/ai100/quiz-w8/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: 500
data_augmentation_options {
random_horizontal_flip {
}
}
data_augmentation_options {
ssd_random_crop {
}
}
}
train_input_reader: {
tf_record_input_reader {
input_path: "/home/hadoop/檢測模型的訓練和使用/data/pet_train.record"
}
label_map_path: "/home/hadoop/檢測模型的訓練和使用/data/labels_items.txt"
}
eval_config: {
num_examples: 47
# Note: The below line limits the evaluation process to 10 evaluations.
# Remove the below line to evaluate indefinitely.
max_evals: 1
}
eval_input_reader: {
tf_record_input_reader {
input_path: "/home/hadoop/檢測模型的訓練和使用/data/pet_val.record"
}
label_map_path: "/home/hadoop/檢測模型的訓練和使用/data/labels_items.txt"
shuffle: true
num_readers: 1
}
配置文件注意文件的路徑要正確✔
爲什麼我們要使用config文件呢?記得之前做分類的項目中我們會在命令行上面輸入很多個參數,而現在在目標檢測中,我們會使用更更多的參數,這些參數可以在config文件中找找到。
這麼多的參數如果是在命令行中執行那麼會多到命令行無法解析,而且我們輸入這麼多的參數更有可能會出先錯誤。
開始訓練
python3 models/research/object_detection/train.py\
--train_dir=./train_dir\
--pipeline_config_path=./data/ssd_mobilenet_v1_pets.config \
--save_interval_secs=300 \
--save_summaries_secs=2
Tensorboard
可以看到我們訓練了1500steps。損失從25+降到了4.5左右。
驗證
我們模型的準確率爲0.9