# copyright (c) 2019 paddlepaddle authors. all rights reserved. # # licensed under the apache license, version 2.0 (the "license"); # you may not use this file except in compliance with the license. # you may obtain a copy of the license at # # http://www.apache.org/licenses/license-2.0 # # unless required by applicable law or agreed to in writing, software # distributed under the license is distributed on an "as is" basis, # without warranties or conditions of any kind, either express or implied. # see the license for the specific language governing permissions and # limitations under the license. import argparse import logging import os import struct import sys import time import unittest import numpy as np import paddle from paddle.base.framework import IrGraph from paddle.framework import core from paddle.static.quantization import QuantInt8OnednnPass paddle.enable_static() logging.basicConfig(format='%(asctime)s-%(levelname)s: %(message)s') _logger = logging.getLogger(__name__) _logger.setLevel(logging.INFO) def parse_args(): parser = argparse.ArgumentParser() parser.add_argument('--batch_size', type=int, default=1, help='Batch size.') parser.add_argument( '--skip_batch_num', type=int, default=0, help='Number of the first minibatches to skip in performance statistics.', ) parser.add_argument( '--debug', action='store_true', help='If used, the graph of Quant model is drawn.', ) parser.add_argument( '--quant_model', type=str, default='', help='A path to a Quant model.' ) parser.add_argument('--infer_data', type=str, default='', help='Data file.') parser.add_argument( '--batch_num', type=int, default=0, help='Number of batches to process. 0 or less means whole dataset. Default: 0.', ) parser.add_argument( '--acc_diff_threshold', type=float, default=0.01, help='Accepted accuracy difference threshold.', ) test_args, args = parser.parse_known_args(namespace=unittest) return test_args, sys.argv[:1] + args class QuantInt8ImageClassificationComparisonTest(unittest.TestCase): """ Test for accuracy comparison of Quant FP32 and INT8 Image Classification inference. """ def _reader_creator(self, data_file='data.bin'): def reader(): with open(data_file, 'rb') as fp: num = fp.read(8) num = struct.unpack('q', num)[0] imgs_offset = 8 img_ch = 3 img_w = 224 img_h = 224 img_pixel_size = 4 img_size = img_ch * img_h * img_w * img_pixel_size label_size = 8 labels_offset = imgs_offset + num * img_size step = 0 while step < num: fp.seek(imgs_offset + img_size * step) img = fp.read(img_size) img = struct.unpack_from(f'{img_ch * img_w * img_h}f', img) img = np.array(img) img.shape = (img_ch, img_w, img_h) fp.seek(labels_offset + label_size * step) label = fp.read(label_size) label = struct.unpack('q', label)[0] yield img, int(label) step += 1 return reader def _get_batch_accuracy(self, batch_output=None, labels=None): total = 0 correct = 0 correct_5 = 0 for n, result in enumerate(batch_output): index = result.argsort() top_1_index = index[-1] top_5_index = index[-5:] total += 1 if top_1_index == labels[n]: correct += 1 if labels[n] in top_5_index: correct_5 += 1 acc1 = float(correct) / float(total) acc5 = float(correct_5) / float(total) return acc1, acc5 def _prepare_for_fp32_onednn(self, graph): ops = graph.all_op_nodes() for op_node in ops: name = op_node.name() if name in ['depthwise_conv2d']: input_var_node = graph._find_node_by_name( op_node.inputs, op_node.input("Input")[0] ) weight_var_node = graph._find_node_by_name( op_node.inputs, op_node.input("Filter")[0] ) output_var_node = graph._find_node_by_name( graph.all_var_nodes(), op_node.output("Output")[0] ) attrs = { name: op_node.op().attr(name) for name in op_node.op().attr_names() } conv_op_node = graph.create_op_node( op_type='conv2d', attrs=attrs, inputs={'Input': input_var_node, 'Filter': weight_var_node}, outputs={'Output': output_var_node}, ) graph.link_to(input_var_node, conv_op_node) graph.link_to(weight_var_node, conv_op_node) graph.link_to(conv_op_node, output_var_node) graph.safe_remove_nodes(op_node) return graph def _predict( self, test_reader=None, model_path=None, batch_size=1, batch_num=1, skip_batch_num=0, transform_to_int8=False, ): place = paddle.CPUPlace() exe = paddle.static.Executor(place) inference_scope = paddle.static.global_scope() with paddle.static.scope_guard(inference_scope): if os.path.exists(os.path.join(model_path, '__model__')): [ inference_program, feed_target_names, fetch_targets, ] = paddle.static.io.load_inference_model( model_path, exe, model_filename=None, params_filename=None ) else: [ inference_program, feed_target_names, fetch_targets, ] = paddle.static.load_inference_model( model_path, exe, model_filename='model', params_filename='params', ) graph = IrGraph(core.Graph(inference_program.desc), for_test=True) if self._debug: graph.draw('.', 'quant_orig', graph.all_op_nodes()) if transform_to_int8: onednn_int8_pass = QuantInt8OnednnPass( _scope=inference_scope, _place=place ) graph = onednn_int8_pass.apply(graph) else: graph = self._prepare_for_fp32_onednn(graph) inference_program = graph.to_program() dshape = [3, 224, 224] outputs = [] infer_accs1 = [] infer_accs5 = [] fpses = [] batch_times = [] total_samples = 0 iters = 0 infer_start_time = time.time() for data in test_reader(): if batch_num > 0 and iters >= batch_num: break if iters == skip_batch_num: total_samples = 0 infer_start_time = time.time() images = [x[0].reshape(dshape) for x in data] images = np.array(images).astype('float32') labels = np.array([x[1] for x in data]).astype('int64') start = time.time() out = exe.run( inference_program, feed={feed_target_names[0]: images}, fetch_list=fetch_targets, ) batch_time = (time.time() - start) * 1000 # in milliseconds outputs.append(out[0]) batch_acc1, batch_acc5 = self._get_batch_accuracy( out[0], labels ) infer_accs1.append(batch_acc1) infer_accs5.append(batch_acc5) samples = len(data) total_samples += samples batch_times.append(batch_time) fps = samples / batch_time * 1000 fpses.append(fps) iters += 1 appx = ' (warm-up)' if iters <= skip_batch_num else '' _logger.info( f'batch {iters}{appx}, acc1: {batch_acc1:.4f}, acc5: {batch_acc5:.4f}, ' f'latency: {batch_time / batch_size:.4f} ms, fps: {fps:.2f}' ) # Postprocess benchmark data batch_latencies = batch_times[skip_batch_num:] batch_latency_avg = np.average(batch_latencies) latency_avg = batch_latency_avg / batch_size fpses = fpses[skip_batch_num:] fps_avg = np.average(fpses) infer_total_time = time.time() - infer_start_time acc1_avg = np.mean(infer_accs1) acc5_avg = np.mean(infer_accs5) _logger.info(f'Total inference run time: {infer_total_time:.2f} s') return outputs, acc1_avg, acc5_avg, fps_avg, latency_avg def _summarize_performance(self, fp32_fps, fp32_lat, int8_fps, int8_lat): _logger.info('--- Performance summary ---') _logger.info( f'FP32: avg fps: {fp32_fps:.2f}, avg latency: {fp32_lat:.4f} ms' ) _logger.info( f'INT8: avg fps: {int8_fps:.2f}, avg latency: {int8_lat:.4f} ms' ) def _compare_accuracy( self, fp32_acc1, fp32_acc5, int8_acc1, int8_acc5, threshold ): _logger.info('--- Accuracy summary ---') _logger.info( f'Accepted top1 accuracy drop threshold: {threshold}. (condition: (FP32_top1_acc - IN8_top1_acc) <= threshold)' ) _logger.info( f'FP32: avg top1 accuracy: {fp32_acc1:.4f}, avg top5 accuracy: {fp32_acc5:.4f}' ) _logger.info( f'INT8: avg top1 accuracy: {int8_acc1:.4f}, avg top5 accuracy: {int8_acc5:.4f}' ) assert fp32_acc1 > 0.0 assert int8_acc1 > 0.0 assert fp32_acc1 - int8_acc1 <= threshold def test_graph_transformation(self): if not core.is_compiled_with_onednn(): return quant_model_path = test_case_args.quant_model assert quant_model_path, ( 'The Quant model path cannot be empty. Please, use the --quant_model option.' ) data_path = test_case_args.infer_data assert data_path, ( 'The dataset path cannot be empty. Please, use the --infer_data option.' ) batch_size = test_case_args.batch_size batch_num = test_case_args.batch_num skip_batch_num = test_case_args.skip_batch_num acc_diff_threshold = test_case_args.acc_diff_threshold self._debug = test_case_args.debug _logger.info('Quant FP32 & INT8 prediction run.') _logger.info(f'Quant model: {quant_model_path}') _logger.info(f'Dataset: {data_path}') _logger.info(f'Batch size: {batch_size}') _logger.info(f'Batch number: {batch_num}') _logger.info(f'Accuracy drop threshold: {acc_diff_threshold}.') _logger.info('--- Quant FP32 prediction start ---') val_reader = paddle.batch( self._reader_creator(data_path), batch_size=batch_size ) fp32_output, fp32_acc1, fp32_acc5, fp32_fps, fp32_lat = self._predict( val_reader, quant_model_path, batch_size, batch_num, skip_batch_num, transform_to_int8=False, ) _logger.info('--- Quant INT8 prediction start ---') val_reader = paddle.batch( self._reader_creator(data_path), batch_size=batch_size ) int8_output, int8_acc1, int8_acc5, int8_fps, int8_lat = self._predict( val_reader, quant_model_path, batch_size, batch_num, skip_batch_num, transform_to_int8=True, ) self._summarize_performance(fp32_fps, fp32_lat, int8_fps, int8_lat) self._compare_accuracy( fp32_acc1, fp32_acc5, int8_acc1, int8_acc5, acc_diff_threshold ) if __name__ == '__main__': global test_case_args test_case_args, remaining_args = parse_args() unittest.main(argv=remaining_args)