fix benchmark train's config for GeneralRecognitionV2, and polish TIPC docs

ppcls/data/dataloader/pk_sampler.py

@@ -53,14 +53,14 @@ class PKSampler(DistributedBatchSampler):
             f"PKSampler configs error, sample_per_id({sample_per_id}) must be a divisor of batch_size({batch_size})."
         assert hasattr(self.dataset,
                        "labels"), "Dataset must have labels attribute."
-        self.sample_per_label = sample_per_id
+        self.sample_per_id = sample_per_id
         self.label_dict = defaultdict(list)
         self.sample_method = sample_method
         for idx, label in enumerate(self.dataset.labels):
             self.label_dict[label].append(idx)
         self.label_list = list(self.label_dict)
-        assert len(self.label_list) * self.sample_per_label > self.batch_size, \
-            "batch size should be smaller than "
+        assert len(self.label_list) * self.sample_per_id >= self.batch_size, \
+            f"batch size({self.batch_size}) should not be bigger than than #classes({len(self.label_list)})*sample_per_id({self.sample_per_id})"
         if self.sample_method == "id_avg_prob":
             self.prob_list = np.array([1 / len(self.label_list)] *
                                       len(self.label_list))
@@ -94,7 +94,7 @@ class PKSampler(DistributedBatchSampler):
                     format(diff))
 
     def __iter__(self):
-        label_per_batch = self.batch_size // self.sample_per_label
+        label_per_batch = self.batch_size // self.sample_per_id
         for _ in range(len(self)):
             batch_index = []
             batch_label_list = np.random.choice(
@@ -104,17 +104,17 @@ class PKSampler(DistributedBatchSampler):
                 p=self.prob_list)
             for label_i in batch_label_list:
                 label_i_indexes = self.label_dict[label_i]
-                if self.sample_per_label <= len(label_i_indexes):
+                if self.sample_per_id <= len(label_i_indexes):
                     batch_index.extend(
                         np.random.choice(
                             label_i_indexes,
-                            size=self.sample_per_label,
+                            size=self.sample_per_id,
                             replace=False))
                 else:
                     batch_index.extend(
                         np.random.choice(
                             label_i_indexes,
-                            size=self.sample_per_label,
+                            size=self.sample_per_id,
                             replace=True))
             if not self.drop_last or len(batch_index) == self.batch_size:
                 yield batch_index

test_tipc/README.md

@@ -115,3 +115,4 @@ bash test_tipc/test_train_inference_python.sh ./test_tipc/configs/MobileNetV3/Mo
 - [test_serving_infer_python 使用](docs/test_serving_infer_python.md)：测试python serving功能。
 - [test_serving_infer_cpp 使用](docs/test_serving_infer_cpp.md)：测试cpp serving功能。
 - [test_train_fleet_inference_python 使用](./docs/test_train_fleet_inference_python.md)：测试基于Python的多机多卡训练与推理等基本功能。
+- [benchmark_train 使用](./docs/benchmark_train.md)：测试基于Python的训练benchmark等基本功能。

test_tipc/configs/GeneralRecognitionV2/GeneralRecognitionV2_PPLCNetV2_base_train_infer_python.txt

@@ -51,7 +51,7 @@ inference:python/predict_rec.py -c configs/inference_rec.yaml
 null:null
 null:null
 ===========================train_benchmark_params==========================
-batch_size:256
+batch_size:128
 fp_items:fp32|fp16
 epoch:1
 --profiler_options:batch_range=[10,20];state=GPU;tracer_option=Default;profile_path=model.profile