add ssd model

Author: saleem-hadad

.gitignore

@@ -5,5 +5,4 @@
 output
 training
 *.record
-model
 inference_graph/

config/fyp.config

@@ -1,130 +1,178 @@
 model {
-  faster_rcnn {
+  ssd {
     num_classes: 1
     image_resizer {
-      keep_aspect_ratio_resizer {
-        # Raw KITTI images have a resolution of 1242x375, if we wish to resize
-        # them to have a height of 600 then their width should be
-        # 1242/(375/600)=1987.2
-        min_dimension: 600
-        max_dimension: 1987
+      fixed_shape_resizer {
+        height: 300
+        width: 300
       }
     }
     feature_extractor {
-      type: 'faster_rcnn_resnet101'
-      first_stage_features_stride: 16
+      type: "ssd_mobilenet_v2"
+      depth_multiplier: 1.0
+      min_depth: 16
+      conv_hyperparams {
+        regularizer {
+          l2_regularizer {
+            weight: 3.99999989895e-05
+          }
+        }
+        initializer {
+          truncated_normal_initializer {
+            mean: 0.0
+            stddev: 0.0299999993294
+          }
+        }
+        activation: RELU_6
+        batch_norm {
+          decay: 0.999700009823
+          center: true
+          scale: true
+          epsilon: 0.0010000000475
+          train: true
+        }
+      }
     }
-    first_stage_anchor_generator {
-      grid_anchor_generator {
-        scales: [0.25, 0.5, 1.0, 2.0]
-        aspect_ratios: [0.5, 1.0, 2.0]
-        height_stride: 16
-        width_stride: 16
+    box_coder {
+      faster_rcnn_box_coder {
+        y_scale: 10.0
+        x_scale: 10.0
+        height_scale: 5.0
+        width_scale: 5.0
       }
     }
-    first_stage_box_predictor_conv_hyperparams {
-      op: CONV
-      regularizer {
-        l2_regularizer {
-          weight: 0.0
-        }
+    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
       }
-      initializer {
-        truncated_normal_initializer {
-          stddev: 0.01
-        }
+    }
+    similarity_calculator {
+      iou_similarity {
       }
     }
-    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
-        fc_hyperparams {
-          op: FC
+    box_predictor {
+      convolutional_box_predictor {
+        conv_hyperparams {
           regularizer {
             l2_regularizer {
-              weight: 0.0
+              weight: 3.99999989895e-05
             }
           }
           initializer {
-            variance_scaling_initializer {
-              factor: 1.0
-              uniform: true
-              mode: FAN_AVG
+            truncated_normal_initializer {
+              mean: 0.0
+              stddev: 0.0299999993294
             }
           }
+          activation: RELU_6
+          batch_norm {
+            decay: 0.999700009823
+            center: true
+            scale: true
+            epsilon: 0.0010000000475
+            train: true
+          }
         }
+        min_depth: 0
+        max_depth: 0
+        num_layers_before_predictor: 0
+        use_dropout: false
+        dropout_keep_probability: 0.800000011921
+        kernel_size: 1
+        box_code_size: 4
+        apply_sigmoid_to_scores: false
       }
     }
-    second_stage_post_processing {
+    anchor_generator {
+      ssd_anchor_generator {
+        num_layers: 6
+        min_scale: 0.20000000298
+        max_scale: 0.949999988079
+        aspect_ratios: 1.0
+        aspect_ratios: 2.0
+        aspect_ratios: 0.5
+        aspect_ratios: 3.0
+        aspect_ratios: 0.333299994469
+      }
+    }
+    post_processing {
       batch_non_max_suppression {
-        score_threshold: 0.0
-        iou_threshold: 0.6
+        score_threshold: 9.99999993923e-09
+        iou_threshold: 0.649999976158
         max_detections_per_class: 100
-        max_total_detections: 300
+        max_total_detections: 100
       }
-      score_converter: SOFTMAX
+      score_converter: SIGMOID
     }
-    second_stage_localization_loss_weight: 2.0
-    second_stage_classification_loss_weight: 1.0
-  }
-}
-
-train_config: {
-  batch_size: 8
-  optimizer {
-    momentum_optimizer: {
-      learning_rate: {
-        manual_step_learning_rate {
-          initial_learning_rate: 0.005
-          schedule {
-            step: 500000
-            learning_rate: .00001
-          }
-          schedule {
-            step: 700000
-            learning_rate: .000001
-          }
+    normalize_loss_by_num_matches: true
+    loss {
+      localization_loss {
+        weighted_smooth_l1 {
+        }
+      }
+      classification_loss {
+        weighted_sigmoid {
         }
       }
-      momentum_optimizer_value: 0.9
+      hard_example_miner {
+        num_hard_examples: 3000
+        iou_threshold: 0.990000009537
+        loss_type: CLASSIFICATION
+        max_negatives_per_positive: 3
+        min_negatives_per_image: 3
+      }
+      classification_weight: 1.0
+      localization_weight: 1.0
     }
-    use_moving_average: false
   }
-  gradient_clipping_by_norm: 10.0
-  fine_tune_checkpoint: "brain/model/model.ckpt"
-  from_detection_checkpoint: true
-  num_steps: 800000
+}
+train_config {
+  batch_size: 128
   data_augmentation_options {
     random_horizontal_flip {
     }
   }
+  data_augmentation_options {
+    ssd_random_crop {
+    }
+  }
+  optimizer {
+    rms_prop_optimizer {
+      learning_rate {
+        exponential_decay_learning_rate {
+          initial_learning_rate: 0.01
+          decay_steps: 800720
+          decay_factor: 0.949999988079
+        }
+      }
+      momentum_optimizer_value: 0.899999976158
+      decay: 0.899999976158
+      epsilon: 1.0
+    }
+  }
+  fine_tune_checkpoint: "brain/preTrianedModel/model.ckpt"
+  fine_tune_checkpoint_type: "detection"
 }
-
-train_input_reader: {
+train_input_reader {
   label_map_path: "brain/config/labels.pbtxt"
-  tf_record_input_reader: {
+  tf_record_input_reader {
     input_path: "brain/config/train.record"
   }
 }
-
-eval_config: {
+eval_config {
+  num_visualizations: 20
+  num_examples: 20
   use_moving_averages: false
-  num_examples: 10
-  num_visualizations: 10
 }
-
-eval_input_reader: {
+eval_input_reader {
   label_map_path: "brain/config/labels.pbtxt"
-  tf_record_input_reader: {
+  shuffle: false
+  num_readers: 1
+  tf_record_input_reader {
     input_path: "brain/config/test.record"
   }
 }

model/checkpoint

@@ -0,0 +1,2 @@
+model_checkpoint_path: "model.ckpt"
+all_model_checkpoint_paths: "model.ckpt"