Unstable loss, extremely low map

[chenaifang]

The training parameters are set as follows：

!/bin/bash

DATASET_DIR=../tfrecords/train TRAIN_DIR=../train_model/ CHECKPOINT_PATH=../checkpoints/vgg_16.ckpt python ../train_ssd_network.py \ --train_dir=${TRAIN_DIR} \ --dataset_dir=${DATASET_DIR} \ --dataset_name=pascalvoc_2007 \ --dataset_split_name=train \ --model_name=ssd_300_vgg \ --checkpoint_path=${CHECKPOINT_PATH} \ --checkpoint_model_scope=vgg_16 \ --checkpoint_exclude_scopes=ssd_300_vgg/conv6,ssd_300_vgg/conv7,ssd_300_vgg/block8,ssd_300_vgg/block9,ssd_300_vgg/block10,ssd_300_vgg/block11,ssd_300_vgg/block4_box,ssd_300_vgg/block7_box,ssd_300_vgg/block8_box,ssd_300_vgg/block9_box,ssd_300_vgg/block10_box,ssd_300_vgg/block11_box \ --trainable_scopes=ssd_300_vgg/conv6,ssd_300_vgg/conv7,ssd_300_vgg/block8,ssd_300_vgg/block9,ssd_300_vgg/block10,ssd_300_vgg/block11,ssd_300_vgg/block4_box,ssd_300_vgg/block7_box,ssd_300_vgg/block8_box,ssd_300_vgg/block9_box,ssd_300_vgg/block10_box,ssd_300_vgg/block11_box \ --save_summaries_secs=60 \ --save_interval_secs=60 \ --weight_decay=0.0005 \ --optimizer=adam \ --learning_rate=0.01 \ --learning_rate_decay_factor=0.97 \ --batch_size=16 \ --max_number_of_steps=50000 \ --gpu_memory_fraction=0.9

loss:

test result:

[chenaifang]

@balancap Could youplease help me with this problem？

[beibeiZ]

how you deal with this problem?

[navy63]

The　ｓａｍｅ　ｐｒｏｂｌｅｍ，

[chenjjjjuu]

the same

[DeepWu]

hello,do you have solve this problem? can you help me?