[MOT] add mcfairmot ptq and update trainer.py (#5298)

* add mcfiarmot ptq

* update docs

configs/mot/mcfairmot/README.md

@@ -44,6 +44,13 @@ PP-tracking provides an AI studio public project tutorial. Please refer to this
  - MOTA is the average MOTA of 4 catecories in the VisDrone Vehicle dataset, and this dataset is extracted from the VisDrone2019 MOT dataset, here we provide the download [link](https://bj.bcebos.com/v1/paddledet/data/mot/visdrone_mcmot_vehicle.zip).
  - The tracker used in MCFairMOT model here is ByteTracker.
 
+### MCFairMOT off-line quantization results on VisDrone Vehicle val-set
+|    Model      |  Compression Strategy | Prediction Delay（T4） |Prediction Delay（V100）| Model Configuration File |Compression Algorithm Configuration File |
+| :--------------| :------- | :------: | :----: | :----: | :----: |
+| DLA-34         | baseline |    41.3  |    21.9 |[Configuration File](./mcfairmot_dla34_30e_1088x608_visdrone_vehicle_bytetracker.yml)|    -     |
+| DLA-34         | off-line quantization   |  37.8    |  21.2  |[Configuration File](./mcfairmot_dla34_30e_1088x608_visdrone_vehicle_bytetracker.yml)|[Configuration File](https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.3/configs/slim/post_quant/mcfairmot_ptq.yml)|
+
+
 ## Getting Start
 
 ### 1. Training
@@ -95,6 +102,14 @@ python deploy/pptracking/python/mot_jde_infer.py --model_dir=output_inference/mc
  - The tracking model is used to predict the video, and does not support the prediction of a single image. The visualization video of the tracking results is saved by default. You can add `--save_mot_txts` to save the txt result file, or `--save_images` to save the visualization images.
  - Each line of the tracking results txt file is `frame,id,x1,y1,w,h,score,cls_id,-1,-1`.
 
+### 6. Off-line quantization
+
+The offline quantization model is calibrated using the VisDrone Vehicle val-set, running as:
+```bash
+CUDA_VISIBLE_DEVICES=0 python3.7 tools/post_quant.py -c configs/mot/mcfairmot/mcfairmot_dla34_30e_1088x608_visdrone_vehicle_bytetracker.yml --slim_config=configs/slim/post_quant/mcfairmot_ptq.yml
+```
+**Notes:**
+ - Offline quantization uses the VisDrone Vehicle val-set dataset and a 4-class vehicle tracking model by default.
 
 ## Citations
 ```

configs/mot/mcfairmot/README_cn.md

@@ -43,6 +43,12 @@ PP-Tracking 提供了AI Studio公开项目案例，教程请参考[PP-Tracking
  - MOTA是VisDrone Vehicle数据集4类车辆目标的平均MOTA, 该数据集是VisDrone数据集中抽出4类车辆类别组成的，此处提供数据集[下载链接](https://bj.bcebos.com/v1/paddledet/data/mot/visdrone_mcmot_vehicle.zip)。
  - MCFairMOT模型此处使用的跟踪器是使用的ByteTracker。
 
+### MCFairMOT 在VisDrone Vehicle val-set上离线量化结果
+|    骨干网络      |  压缩策略 | 预测时延（T4） |预测时延（V100）| 配置文件 |压缩算法配置文件 |
+| :--------------| :------- | :------: | :----: | :----: | :----: |
+| DLA-34         | baseline |    41.3  |    21.9 |[配置文件](./mcfairmot_dla34_30e_1088x608_visdrone_vehicle_bytetracker.yml)|    -     |
+| DLA-34         | 离线量化   |  37.8    |  21.2  |[配置文件](./mcfairmot_dla34_30e_1088x608_visdrone_vehicle_bytetracker.yml)|[配置文件](https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.3/configs/slim/post_quant/mcfairmot_ptq.yml)|
+
 ## 快速开始
 
 ### 1. 训练
@@ -93,6 +99,14 @@ python deploy/pptracking/python/mot_jde_infer.py --model_dir=output_inference/mc
  - 跟踪模型是对视频进行预测，不支持单张图的预测，默认保存跟踪结果可视化后的视频，可添加`--save_mot_txts`表示保存跟踪结果的txt文件，或`--save_images`表示保存跟踪结果可视化图片。
  - 多类别跟踪结果txt文件每行信息是`frame,id,x1,y1,w,h,score,cls_id,-1,-1`。
 
+### 6. 离线量化
+
+使用 VisDrone Vehicle val-set 对离线量化模型进行校准，运行方式：
+```bash
+CUDA_VISIBLE_DEVICES=0 python3.7 tools/post_quant.py -c configs/mot/mcfairmot/mcfairmot_dla34_30e_1088x608_visdrone_vehicle_bytetracker.yml --slim_config=configs/slim/post_quant/mcfairmot_ptq.yml
+```
+**注意:**
+ - 离线量化默认使用的是VisDrone Vehicle val-set数据集以及4类车辆跟踪模型。
 
 ## 引用
 ```

configs/slim/post_quant/mcfairmot_ptq.yml

+weights:  https://paddledet.bj.bcebos.com/models/mot/mcfairmot_dla34_30e_1088x608_visdrone_vehicle_bytetracker.pdparams
+slim: PTQ
+
+PTQ:
+  ptq_config: {
+      'activation_quantizer': 'HistQuantizer',
+      'upsample_bins': 127,
+      'hist_percent': 0.999}
+  quant_batch_num: 10
+  fuse: True

ppdet/engine/trainer.py

@@ -73,6 +73,10 @@ class Trainer(object):
             logger.error('DeepSORT has no need of training on mot dataset.')
             sys.exit(1)
 
+        if cfg.architecture == 'FairMOT' and self.mode == 'eval':
+            images = self.parse_mot_images(cfg)
+            self.dataset.set_images(images)
+
         if self.mode == 'train':
             self.loader = create('{}Reader'.format(self.mode.capitalize()))(
                 self.dataset, cfg.worker_num)
@@ -114,14 +118,17 @@ class Trainer(object):
         # EvalDataset build with BatchSampler to evaluate in single device
         # TODO: multi-device evaluate
         if self.mode == 'eval':
-            self._eval_batch_sampler = paddle.io.BatchSampler(
-                self.dataset, batch_size=self.cfg.EvalReader['batch_size'])
-            reader_name = '{}Reader'.format(self.mode.capitalize())
-            # If metric is VOC, need to be set collate_batch=False.
-            if cfg.metric == 'VOC':
-                cfg[reader_name]['collate_batch'] = False
-            self.loader = create(reader_name)(self.dataset, cfg.worker_num,
-                                              self._eval_batch_sampler)
+            if cfg.architecture == 'FairMOT':
+                self.loader = create('EvalMOTReader')(self.dataset, 0)
+            else:
+                self._eval_batch_sampler = paddle.io.BatchSampler(
+                    self.dataset, batch_size=self.cfg.EvalReader['batch_size'])
+                reader_name = '{}Reader'.format(self.mode.capitalize())
+                # If metric is VOC, need to be set collate_batch=False.
+                if cfg.metric == 'VOC':
+                    cfg[reader_name]['collate_batch'] = False
+                self.loader = create(reader_name)(self.dataset, cfg.worker_num,
+                                                  self._eval_batch_sampler)
         # TestDataset build after user set images, skip loader creation here
 
         # build optimizer in train mode
@@ -759,3 +766,28 @@ class Trainer(object):
         flops = flops(self.model, input_spec) / (1000**3)
         logger.info(" Model FLOPs : {:.6f}G. (image shape is {})".format(
             flops, input_data['image'][0].unsqueeze(0).shape))
+
+    def parse_mot_images(self, cfg):
+        import glob
+        # for quant
+        dataset_dir = cfg['EvalMOTDataset'].dataset_dir
+        data_root = cfg['EvalMOTDataset'].data_root
+        data_root = '{}/{}'.format(dataset_dir, data_root)
+        seqs = os.listdir(data_root)
+        seqs.sort()
+        all_images = []
+        for seq in seqs:
+            infer_dir = os.path.join(data_root, seq)
+            assert infer_dir is None or os.path.isdir(infer_dir), \
+                "{} is not a directory".format(infer_dir)
+            images = set()
+            exts = ['jpg', 'jpeg', 'png', 'bmp']
+            exts += [ext.upper() for ext in exts]
+            for ext in exts:
+                images.update(glob.glob('{}/*.{}'.format(infer_dir, ext)))
+            images = list(images)
+            images.sort()
+            assert len(images) > 0, "no image found in {}".format(infer_dir)
+            all_images.extend(images)
+            logger.info("Found {} inference images in total.".format(len(images)))
+        return all_images
\ No newline at end of file