Merge pull request #16 from Oneflow-Inc/dev_test_map

replace depreciated function with new

oneflow_yolov3/model/predict_with_print_box.py

@@ -66,7 +66,7 @@ args = parser.parse_args()
 
 flow.config.load_library(oneflow_yolov3.lib_path())
 func_config = flow.FunctionConfig()
-func_config.default_distribute_strategy(flow.scope.consistent_view())
+func_config.default_logical_view(flow.scope.consistent_view())
 func_config.default_data_type(flow.float)
 if args.use_tensorrt != 0:
     func_config.use_tensorrt(True)

oneflow_yolov3/model/yolo_evaluate.py

@@ -4,6 +4,7 @@ import os
 
 import numpy as np
 import oneflow as flow
+import oneflow.typing as tp
 import utils
 from data_preprocess import batch_image_preprocess_with_label
 from tqdm import tqdm
@@ -54,38 +55,22 @@ parser.add_argument(
     default=0,
     required=False)
 parser.add_argument("-image_paths", "--image_paths", type=str, required=False)
-
-
 args = parser.parse_args()
 
 
 flow.config.load_library(oneflow_yolov3.lib_path())
 func_config = flow.FunctionConfig()
-func_config.default_distribute_strategy(flow.scope.consistent_view())
+func_config.default_logical_view(flow.scope.consistent_view())
 func_config.default_data_type(flow.float)
 if args.use_tensorrt != 0:
     func_config.use_tensorrt(True)
 # func_config.tensorrt.use_fp16()
 
 
-input_blob_def_dict = {
-    "images": flow.FixedTensorDef(
-        (args.batch_size,
-         3,
-         args.image_height,
-         args.image_width),
-        dtype=flow.float),
-    "origin_image_info": flow.FixedTensorDef(
-        (args.batch_size,
-         2),
-        dtype=flow.int32),
-}
-
-
 @flow.global_function(func_config)
-def yolo_user_op_eval_job(
-        images=input_blob_def_dict["images"],
-        origin_image_info=input_blob_def_dict["origin_image_info"]):
+def yolo_user_op_eval_job(images:tp.Numpy.Placeholder((args.batch_size, 3, args.image_height, args.image_width), dtype=flow.float),
+                          origin_image_info:tp.Numpy.Placeholder((args.batch_size, 2), dtype=flow.int32)
+                          ):
     yolo_pos_result, yolo_prob_result = YoloPredictNet(
         images, origin_image_info, trainable=False)
     return yolo_pos_result, yolo_prob_result, origin_image_info
@@ -100,7 +85,7 @@ if __name__ == "__main__":
         names = f.read().split('\n')
     names = list(filter(None, names))
     flow.config.gpu_device_num(args.gpu_num_per_node)
-    # load model
+    # Load model
     check_point = flow.train.CheckPoint()
     check_point.load(args.model_load_dir)
 
@@ -110,6 +95,8 @@ if __name__ == "__main__":
             path_list.append(line.strip('\n'))
 
     iter_num = math.floor(len(path_list) / float(args.batch_size))
+
+    # evaluate mAP
     """
     reference:
     https://github.com/ultralytics/yolov3/blob/master/test.py

oneflow_yolov3/model/yolo_train.py

@@ -97,7 +97,7 @@ args = parser.parse_args()
 flow.config.gpu_device_num(args.gpu_num_per_node)
 flow.config.load_library(oneflow_yolov3.lib_path())
 func_config = flow.FunctionConfig()
-func_config.default_distribute_strategy(flow.scope.consistent_view())
+func_config.default_logical_view(flow.scope.consistent_view())
 func_config.default_data_type(flow.float)
 func_config.train.primary_lr(args.base_lr)
 func_config.train.model_update_conf(dict(naive_conf={}))

yolo_train.sh

@@ -9,5 +9,5 @@ python3 oneflow_yolov3/model/yolo_train.py \
 --num_epoch=130 --model_load_dir=$model_dir \
 --classes=80 --num_boxes=90 --save_frequency=1 \
 --model_save_dir="save_model" \
---dataset_dir="data/COCO/test_trainvalno5k.txt"
-# --dataset_dir="data/trainvalno5k.txt"
+--dataset_dir="data/test_trainvalno5k.txt"
+# --dataset_dir="data/COCO/trainvalno5k.txt"