search

tensorflow / models

Models and examples built with TensorFlow
Other
76.97k stars 45.79k forks source link

SSD-MobileNetV3 Large and SSD-MobileDet-CPU not restoring the checkpoint provided, and starting from scratch #9284

Open RomStriker opened 3 years ago

RomStriker commented 3 years ago

Prerequisites

Please answer the following questions for yourself before submitting an issue.

1. The entire URL of the file you are using

https://github.com/tensorflow/models/blob/master/research/object_detection/model_main.py

2. Describe the bug

I am fine-tuning SSD-MobileNetV3 Large and SSD-MobileDet-CPU on the COCO 2017 dataset but with only book classes. I have created a new dataset for this and inspected the dataset and it is good. I have also modified the config file to my needs. When I start the training, it just ignores the fine_tune_checkpoint provided in the config file and starts from scratch. However, if I do the same process but with the checkpoint in the model_dir directory instead, it tries to restore it but since I have different number of classes, it gives an error. How can make the training process restore the checkpoint properly? I also tried with normal COCO dataset with all 90 classes, and when I start the training, fine_tune_checkpoint is ignore, but if I put the checkpoint in the model_dir, it is restored properly.

3. Steps to reproduce

Config file:

# SSDLite with MobileDet-CPU feature extractor.
# Reference: Xiong & Liu et al., https://arxiv.org/abs/2004.14525
# Trained on COCO, initialized from scratch.
#
# 0.45B MulAdds, 4.21M Parameters. Latency is 113ms on Pixel 1 CPU.
# Achieves 24.0 mAP on COCO14 minival dataset.
# Achieves 23.5 mAP on COCO17 val dataset.
#
# This config is TPU compatible.

model {
  ssd {
    inplace_batchnorm_update: true
    freeze_batchnorm: false
    num_classes: 1
    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
        use_matmul_gather: true
      }
    }
    similarity_calculator {
      iou_similarity {
      }
    }
    encode_background_as_zeros: true
    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: 320
        width: 320
      }
    }
    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: 3
        use_depthwise: true
        box_code_size: 4
        apply_sigmoid_to_scores: false
        class_prediction_bias_init: -4.6
        conv_hyperparams {
          activation: RELU_6,
          regularizer {
            l2_regularizer {
              weight: 0.00004
            }
          }
          initializer {
            random_normal_initializer {
              stddev: 0.03
              mean: 0.0
            }
          }
          batch_norm {
            train: true,
            scale: true,
            center: true,
            decay: 0.97,
            epsilon: 0.001,
          }
        }
      }
    }
    feature_extractor {
      type: 'ssd_mobiledet_cpu'
      min_depth: 16
      depth_multiplier: 1.0
      use_depthwise: true
      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: 0.97,
          epsilon: 0.001,
        }
      }
      override_base_feature_extractor_hyperparams: false
    }
    loss {
      classification_loss {
        weighted_sigmoid_focal {
          alpha: 0.75,
          gamma: 2.0
        }
      }
      localization_loss {
        weighted_smooth_l1 {
          delta: 1.0
        }
      }
      classification_weight: 1.0
      localization_weight: 1.0
    }
    normalize_loss_by_num_matches: true
    normalize_loc_loss_by_codesize: true
    post_processing {
      batch_non_max_suppression {
        score_threshold: 1e-8
        iou_threshold: 0.6
        max_detections_per_class: 100
        max_total_detections: 100
        use_static_shapes: true
      }
      score_converter: SIGMOID
    }
  }
}

train_config: {
  batch_size: 64
  sync_replicas: true
  startup_delay_steps: 0
  replicas_to_aggregate: 1
  num_steps: 800000
  data_augmentation_options {
    random_horizontal_flip {
    }
  }
  data_augmentation_options {
    ssd_random_crop {
    }
  }
  optimizer {
    momentum_optimizer: {
      learning_rate: {
        cosine_decay_learning_rate {
          learning_rate_base: 0.8
          total_steps: 800000
          warmup_learning_rate: 0.13333
          warmup_steps: 100
        }
      }
      momentum_optimizer_value: 0.9
    }
    use_moving_average: false
  }
  fine_tune_checkpoint: "./checkpoints/model.ckpt-400000"
  max_number_of_boxes: 100
  unpad_groundtruth_tensors: false
  fine_tune_checkpoint_type: "detection"
  fine_tune_checkpoint_version: V1
  load_all_detection_checkpoint_vars: true
}

train_input_reader: {
  tf_record_input_reader {
    input_path: "./tf_record_coco_books/coco_train.record"
  }
  label_map_path: "./tf_record_coco_books/label_map.pbtxt"
}

eval_config: {
  metrics_set: "coco_detection_metrics"
  use_moving_averages: false
}

eval_input_reader: {
  tf_record_input_reader {
    input_path: "./tf_record_coco_books/coco_val.record"
  }
  label_map_path: "./tf_record_coco_books/label_map.pbtxt"
  shuffle: false
  num_readers: 1
}

4. Expected behavior

N/A

5. Additional context

N/A

6. System information

RomStriker commented 3 years ago

The issue rises from line 446 in model_lib.py load_pretrained = hparams.load_pretrained if hparams else False; because one of the previous commits changed hparams to None as mention in #8695 so load_pretrained is always False. Setting it to True, and reinstalling the object_detection library fixes the problem. However, as mentioned in #8695 there should be a better fix for this.

mohammedayub44 commented 3 years ago

@RomStriker I run into a somewhat similar issue while finetuning from available checkpoints downloaded form Model Zoo. Since the model zoo provides a trained checkpoint in a checkpoint directory, passing this directory gives an error. Training script throws warnings that checkpoint directory is ignored and tried to create a checkpoint.tmpXXXXX directory and error out. I'm guessing it's also because of this reason, if it's still not fixed. Any thoughts.

However the training starts fine from scratch if I don't give the checkpoint directory.

copiousOne commented 3 years ago

People, myself included, trust TensorFlow V1 over V2. I am so close to switching libraries. Any suggestions? Does PyTorch work well with Horovod or scale easily to multi-GPU training?