Fix distiller (#1049)

* fix distiller

* fix distiller

* fix distiller

* demo imagenet

* demo imagenet

demo/auto-compression/demo_imagenet.py

@@ -25,9 +25,10 @@ add_arg('save_dir',                    str,    None,         "directory to save
 add_arg('devices',                     str,    'gpu',        "which device used to compress.")
 add_arg('batch_size',                  int,    1,            "train batch size.")
 add_arg('config_path',                 str,    None,         "path of compression strategy config.")
-# yapf: enable
+add_arg('data_dir',                    str,    None,         "path of dataset")
 
 
+# yapf: enable
 def reader_wrapper(reader):
     def gen():
         for i, data in enumerate(reader()):
@@ -38,7 +39,8 @@ def reader_wrapper(reader):
 
 
 def eval_function(exe, compiled_test_program, test_feed_names, test_fetch_list):
-    val_reader = paddle.batch(reader.val(), batch_size=1)
+
+    val_reader = paddle.batch(reader.val(data_dir=data_dir), batch_size=1)
     image = paddle.static.data(
         name='x', shape=[None, 3, 224, 224], dtype='float32')
     label = paddle.static.data(name='label', shape=[None, 1], dtype='int64')
@@ -47,7 +49,6 @@ def eval_function(exe, compiled_test_program, test_feed_names, test_fetch_list):
     for batch_id, data in enumerate(val_reader()):
         # top1_acc, top5_acc
         if len(test_feed_names) == 1:
-            # eval "infer model", which input is image, output is classification probability
             image = data[0][0].reshape((1, 3, 224, 224))
             label = [[d[1]] for d in data]
             pred = exe.run(compiled_test_program,
@@ -76,6 +77,8 @@ def eval_function(exe, compiled_test_program, test_feed_names, test_fetch_list):
                 fetch_list=test_fetch_list)
             result = [np.mean(r) for r in result]
             results.append(result)
+        if batch_id % 5000 == 0:
+            print('Eval iter: ', batch_id)
     result = np.mean(np.array(results), axis=0)
     return result[0]
 
@@ -85,8 +88,10 @@ if __name__ == '__main__':
     print_arguments(args)
     paddle.enable_static()
     compress_config, train_config = load_config(args.config_path)
+    data_dir = args.data_dir
 
-    train_reader = paddle.batch(reader.train(), batch_size=64)
+    train_reader = paddle.batch(
+        reader.train(data_dir=data_dir), batch_size=args.batch_size)
     train_dataloader = reader_wrapper(train_reader)
 
     ac = AutoCompression(

demo/auto-compression/run_imagenet.sh

@@ -5,4 +5,5 @@ python3.7 demo_imagenet.py \
     --save_dir='./save_qat_mbv2/' \
     --devices='cpu' \
     --batch_size=2 \
+    --data_dir='data/ILSVRC2012/' \
     --config_path='./configs/CV/mbv2_ptq_hpo.yaml'

demo/imagenet_reader.py

@@ -187,6 +187,7 @@ def test(data_dir=DATA_DIR):
 class ImageNetDataset(Dataset):
     def __init__(self, data_dir=DATA_DIR, mode='train'):
         super(ImageNetDataset, self).__init__()
+        self.data_dir = data_dir
         train_file_list = os.path.join(data_dir, 'train_list.txt')
         val_file_list = os.path.join(data_dir, 'val_list.txt')
         test_file_list = os.path.join(data_dir, 'test_list.txt')
@@ -204,7 +205,7 @@ class ImageNetDataset(Dataset):
 
     def __getitem__(self, index):
         sample = self.data[index]
-        data_path = os.path.join(DATA_DIR, sample[0])
+        data_path = os.path.join(self.data_dir, sample[0])
         if self.mode == 'train':
             data, label = process_image(
                 [data_path, sample[1]],

paddleslim/auto_compression/compressor.py

@@ -163,6 +163,8 @@ class AutoCompression:
                 self._exe, self._places, config_dict, train_program_info,
                 self._strategy)
 
+
+
         if self.train_config.use_fleet:
             dist_strategy = _prepare_fleet_strategy(self.train_config)
         else:
@@ -188,6 +190,8 @@ class AutoCompression:
 
         self._exe.run(train_program_info.startup_program)
 
+
+
         if (not self.train_config.use_fleet
             ) and self.train_config.amp_config is not None:
             if hasattr(self.train_config.amp_config, 'use_pure_fp16'

paddleslim/auto_compression/create_compressed_program.py

@@ -96,13 +96,18 @@ def _load_program_and_merge(executor,
                             params_filename,
                             teacher_idx=None,
                             feed_target_names=None):
+
+    scope = paddle.static.global_scope()
+    new_scope = paddle.static.Scope()
     try:
+        with paddle.static.scope_guard(new_scope):
             [teacher_program, teacher_feed_target_names, teacher_fetch_targets]= paddle.fluid.io.load_inference_model( \
             dirname=model_dir, \
             model_filename=model_filename, \
             params_filename=params_filename, \
             executor=executor)
     except:
+        with paddle.static.scope_guard(new_scope):
             [teacher_program, teacher_feed_target_names, teacher_fetch_targets]= paddle.static.load_inference_model( \
             path_prefix=model_dir, \
             executor=executor)
@@ -130,6 +135,7 @@ def _load_program_and_merge(executor,
         train_program,
         data_name_map,
         place,
+        teacher_scope=new_scope,
         name_prefix=teacher_name_prefix,
         merge_feed=config.get('merge_feed') or True)
     if teacher_idx == None or teacher_idx == 1:
@@ -280,7 +286,6 @@ def build_quant_program(executor, place, config, train_program_info,
     assert isinstance(config, dict), "quant config must be dict"
     default_config = _quant_config_default
     default_config.update(config)
-    print(default_config)
     config = _parse_configs(default_config)
 
     use_pact = config["use_pact"]

paddleslim/dist/single_distiller.py

@@ -22,6 +22,7 @@ def merge(teacher_program,
           data_name_map,
           place,
           scope=None,
+          teacher_scope=None,
           name_prefix='teacher_',
           merge_feed=True):
     """Merge teacher program into student program and add a uniform prefix to the
@@ -48,6 +49,8 @@ def merge(teacher_program,
     """
     if scope == None:
         scope = paddle.static.global_scope()
+    if teacher_scope == None:
+        teacher_scope = scope
     teacher_program = teacher_program.clone(for_test=True)
     for teacher_var in teacher_program.list_vars():
         skip_rename = False
@@ -60,7 +63,7 @@ def merge(teacher_program,
                 new_name = name_prefix + teacher_var.name
             if not skip_rename:
                 # scope var rename
-                old_var = scope.var(teacher_var.name).get_tensor()
+                old_var = teacher_scope.var(teacher_var.name).get_tensor()
                 renamed_var = scope.var(new_name).get_tensor()
                 renamed_var.set(np.array(old_var), place)