Python version -> 3.6.4
TF version -> 1.8.0
Running on : CPU
OS: Windows 10
大家好！我正在尝试从TF对象检测API中微调模型以进行对象细分。我使用预先训练的mask_rcnn_inception_v2_coco_2018_01_28作为基准，并且正在LFW Dataset上进行训练，以从头部，头发和面部的图片中进行分割。
我遍历了不同的教程来生成数据集，显然，使用create_pet_tf_record.py脚本将我的数据集成功转换为Train和eval TFrecord文件（我只是修改了加载XML的部分通过直接生成字典以及所有所需部分的函数的图像文件。
train.py脚本没有引发错误，并且训练似乎已经按预期进行，但是当我尝试在1000步之后评估模型（已将其导出到.pb文件之后）时，使用教程中提供的Jupyter笔记本，该模型甚至没有在脸部周围生成盒子。所以我试图观察张量板上的训练，并且发现在所有训练过程中模板损失都保持为零
此外（我不知道这是否有帮助），我与其他教程有所不同，我无法从张量板上的评估集中查看图像中的预测。
未找到图像数据。
可能的原因： 您尚未将任何图像数据写入事件文件。 TensorBoard找不到您的事件文件。
但是我没有碰到摘要行为。
我的配置文件如下：
# Mask R-CNN with Inception V2
# Configured 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 {
faster_rcnn {
num_classes: 1
image_resizer {
keep_aspect_ratio_resizer {
min_dimension: 220
max_dimension: 1365
}
}
number_of_stages: 3
feature_extractor {
type: 'faster_rcnn_inception_v2'
first_stage_features_stride: 16
}
first_stage_anchor_generator {
grid_anchor_generator {
scales: [0.25, 0.5, 1.0, 2.0]
aspectratios: [0.5, 1.0, 2.0]
height_stride: 16
width_stride: 16
}
}
first_stage_box_predictor_conv_hyperparams {
op: CONV
regularizer {
l2_regularizer {
weight: 0.0
}
}
initializer {
truncated_normal_initializer {
stddev: 0.0
}
}
}
first_stage_nms_score_threshold: 0.0
first_stage_nms_iou_threshold: 0.7
first_stage_max_proposals: 300
first_stage_localization_loss_weight: 2.0
first_stage_objectness_loss_weight: 1.0
initial_crop_size: 14
maxpool_kernel_size: 2
maxpool_stride: 2
second_stage_box_predictor {
mask_rcnn_box_predictor {
use_dropout: false
dropout_keep_probability: 1.0
predict_instance_masks: true
mask_height: 15
mask_width: 15
mask_prediction_conv_depth: 0
mask_prediction_num_conv_layers: 2
fc_hyperparams {
op: FC
regularizer {
l2_regularizer {
weight: 0.1
}
}
initializer {
variance_scaling_initializer {
factor: 1.0
uniform: true
mode: FAN_AVG
}
}
}
conv_hyperparams {
op: CONV
regularizer {
l2_regularizer {
weight: 0.0
}
}
initializer {
truncated_normal_initializer {
stddev: 0.01
}
}
}
}
}
second_stage_post_processing {
batch_non_max_suppression {
score_threshold: 0.0
iou_threshold: 0.6
max_detections_per_class: 100
max_total_detections: 300
}
score_converter: SOFTMAX
}
second_stage_localization_loss_weight: 2.0
second_stage_classification_loss_weight: 1.0
second_stage_mask_prediction_loss_weight: 4.0
}
}
train_config: {
batch_size: 1
optimizer {
momentum_optimizer: {
learning_rate: {
manual_step_learning_rate {
initial_learning_rate: 0.00002
schedule {
step: 1000
learning_rate: .000002
}
schedule {
step: 2000
learning_rate: .0000002
}
}
}
momentum_optimizer_value: 0.9
}
use_moving_average: false
}
gradient_clipping_by_norm: 10.0
fine_tune_checkpoint: "model/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: 200000
data_augmentation_options {
random_horizontal_flip {
}
}
}
train_input_reader: {
tf_record_input_reader {
input_path: "data/pictures_with_masks_train.record-00000-of-00001"
}
label_map_path: "data/label.pbtxt.txt"
load_instance_masks: True
mask_type: PNG_MASKS
}
eval_config: {
num_examples: 878
# 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/pictures_with_masks_val.record-00000-of-00001"
}
label_map_path: "data/label.pbtxt.txt"
load_instance_masks: True
mask_type: PNG_MASKS
shuffle: False
num_readers: 1
}
我将非常感谢您提供的任何帮助，如果您有任何想法，请随时回复！
谢谢！