score:0

Which model do you use? You might try using a different network for training. I had the same problem when used ssd_mobilenet_v1_coco and ssd_mobilenet_v1_coco.config. Switched to ssd_mobilenet_v1_pets.config, which has additional parameter anchorwise_output, and it started to train.

score:1

problem solved. In your pipeline.config make

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
}

to

loss {
classification_loss {
weighted_sigmoid {
anchorwise_output: true #add this
}
}
localization_loss {
weighted_smooth_l1 {
anchorwise_output: true #add this
}
}
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
}

enjoy...