Merge branch 'develop' into pact_clip

paddleslim/quant/quanter.py

@@ -12,7 +12,9 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+import os
 import copy
+import json
 import logging
 
 import paddle
@@ -33,8 +35,7 @@ from ..common import get_logger
 _logger = get_logger(__name__, level=logging.INFO)
 
 WEIGHT_QUANTIZATION_TYPES = [
-    'abs_max', 'channel_wise_abs_max', 'range_abs_max',
-    'moving_average_abs_max'
+    'abs_max', 'channel_wise_abs_max', 'range_abs_max', 'moving_average_abs_max'
 ]
 WEIGHT_QUANTIZATION_TYPES_TENSORRT = ['channel_wise_abs_max']
 
@@ -55,6 +56,8 @@ TENSORRT_OP_TYPES = [
     'leaky_relu'
 ]
 
+VARS_MAPPING_TABLE = './mapping_table_for_saving_inference_model'
+
 _quant_config_default = {
     # weight quantize type, default is 'channel_wise_abs_max'
     'weight_quantize_type': 'channel_wise_abs_max',
@@ -81,6 +84,18 @@ _quant_config_default = {
 }
 
 
+def load_dict():
+    with open(VARS_MAPPING_TABLE, 'r') as file:
+        data = file.read()
+        data = json.loads(data)
+        return data
+
+
+def save_dict(table):
+    with open(VARS_MAPPING_TABLE, 'w') as file:
+        file.write(json.dumps(table))
+
+
 def _parse_configs(user_config):
     """
     check if user's configs are valid.
@@ -267,6 +282,15 @@ def quant_aware(program,
         scope=scope, place=place, moving_rate=config['moving_rate'])
     out_scale_training_pass.apply(main_graph)
 
+    if (weight_preprocess_func is not None or
+            act_preprocess_func is not None) and not for_test:
+        _logger.info(
+            "When a preprocess_func is used in quant_aware, Need to save a mapping table to match variable names in the convert phase."
+        )
+        _logger.info("The mapping table is saved as '{}'.".format(
+            VARS_MAPPING_TABLE))
+        save_dict(main_graph.out_node_mapping_table)
+
     if for_test:
         quant_program = main_graph.to_program()
     else:
@@ -274,27 +298,28 @@ def quant_aware(program,
     return quant_program
 
 
-def quant_post_static(executor,
-               model_dir,
-               quantize_model_path,
-               batch_generator=None,
-               sample_generator=None,
-               model_filename=None,
-               params_filename=None,
-               save_model_filename='__model__',
-               save_params_filename='__params__',
-               batch_size=16,
-               batch_nums=None,
-               scope=None,
-               algo='KL',
-               quantizable_op_type=["conv2d", "depthwise_conv2d", "mul"],
-               is_full_quantize=False,
-               weight_bits=8,
-               activation_bits=8,
-               activation_quantize_type='range_abs_max',
-               weight_quantize_type='channel_wise_abs_max',
-               is_use_cache_file=False,
-               cache_dir="./temp_post_training"):
+def quant_post_static(
+        executor,
+        model_dir,
+        quantize_model_path,
+        batch_generator=None,
+        sample_generator=None,
+        model_filename=None,
+        params_filename=None,
+        save_model_filename='__model__',
+        save_params_filename='__params__',
+        batch_size=16,
+        batch_nums=None,
+        scope=None,
+        algo='KL',
+        quantizable_op_type=["conv2d", "depthwise_conv2d", "mul"],
+        is_full_quantize=False,
+        weight_bits=8,
+        activation_bits=8,
+        activation_quantize_type='range_abs_max',
+        weight_quantize_type='channel_wise_abs_max',
+        is_use_cache_file=False,
+        cache_dir="./temp_post_training"):
     """
     The function utilizes static post training quantization method to
     quantize the fp32 model. It uses calibrate data to calculate the
@@ -381,6 +406,7 @@ def quant_post_static(executor,
         model_filename=save_model_filename,
         params_filename=save_params_filename)
 
+
 # We have changed the quant_post to quant_post_static.
 # For compatibility, we keep quant_post api for now, and it will be
 # deprecated in the future.
@@ -438,6 +464,9 @@ def convert(program, place, config=None, scope=None, save_int8=False):
         activation_bits=config['activation_bits'],
         weight_quantize_type=config['weight_quantize_type'])
 
+    if os.path.exists(VARS_MAPPING_TABLE):
+        test_graph.out_node_mapping_table = load_dict()
+
     freeze_pass.apply(test_graph)
     freezed_program = test_graph.to_program()
 
@@ -451,14 +480,14 @@ def convert(program, place, config=None, scope=None, save_int8=False):
 
 
 def quant_post_dynamic(model_dir,
-                           save_model_dir,
-                           model_filename=None,
-                           params_filename=None,
-                           save_model_filename=None,
-                           save_params_filename=None,
-                           quantizable_op_type=["conv2d", "mul"],
-                           weight_bits=8,
-                           generate_test_model=False):
+                       save_model_dir,
+                       model_filename=None,
+                       params_filename=None,
+                       save_model_filename=None,
+                       save_params_filename=None,
+                       quantizable_op_type=["conv2d", "mul"],
+                       weight_bits=8,
+                       generate_test_model=False):
     '''
     The function utilizes static post training quantization method to
     quantize the fp32 model. In details, it quantizes the weight of some
@@ -517,4 +546,4 @@ def quant_post_dynamic(model_dir,
 # We have changed the quant_post_only_weight to quant_post_dynamic.
 # For compatibility, we keep quant_post_only_weight api for now,
 # and it will be deprecated in the future.
-quant_post_only_weight = quant_post_dynamic
\ No newline at end of file
+quant_post_only_weight = quant_post_dynamic