refactor

1. rm Global.eval_mode
2. add Global.task
3. mv type_name to ppcls.utils
4. build dataloader, loss, metric by mode

ppcls/data/__init__.py

@@ -110,7 +110,7 @@ def build(config, mode, use_dali=False, seed=None):
     config_dataset = copy.deepcopy(config_dataset)
     dataset_name = config_dataset.pop('name')
     if 'batch_transform_ops' in config_dataset:
-        batch_transform = config_dataset.pop('batch_transform_ops')
+        batch_transform = config_dataset['batch_transform_ops']
     else:
         batch_transform = None
 
@@ -254,10 +254,11 @@ def build_dataloader(config, mode):
 
     if mode == "eval" or (mode == "train" and
                           config["Global"]["eval_during_train"]):
-        if config["Global"]["eval_mode"] in ["classification", "adaface"]:
+        task = config["Global"].get("task", "classification")
+        if task in ["classification", "adaface"]:
             dataloader_dict["Eval"] = build(
                 config["DataLoader"], "Eval", use_dali, seed=None)
-        elif config["Global"]["eval_mode"] == "retrieval":
+        elif task == "retrieval":
             if len(config["DataLoader"]["Eval"].keys()) == 1:
                 key = list(config["DataLoader"]["Eval"].keys())[0]
                 dataloader_dict["GalleryQuery"] = build(

ppcls/data/dataloader/dali.py

@@ -42,7 +42,7 @@ from ppcls.data.preprocess.ops.dali_operators import RandomRot90
 from ppcls.data.preprocess.ops.dali_operators import RandomRotation
 from ppcls.data.preprocess.ops.dali_operators import ResizeImage
 from ppcls.data.preprocess.ops.dali_operators import ToCHWImage
-from ppcls.engine.train.utils import type_name
+from ppcls.utils import type_name
 from ppcls.utils import logger
 
 INTERP_MAP = {

ppcls/engine/train/train_progressive.py

@@ -14,8 +14,7 @@
 from __future__ import absolute_import, division, print_function
 
 from ppcls.data import build_dataloader
-from ppcls.engine.train.utils import type_name
-from ppcls.utils import logger
+from ppcls.utils import logger, type_name
 from .regular_train_epoch import regular_train_epoch
 
 

ppcls/engine/train/utils.py

@@ -14,7 +14,7 @@
 from __future__ import absolute_import, division, print_function
 
 import datetime
-from ppcls.utils import logger
+from ppcls.utils import logger, type_name
 from ppcls.utils.misc import AverageMeter
 
 
@@ -75,8 +75,3 @@ def log_info(trainer, batch_size, epoch_id, iter_id):
             value=trainer.output_info[key].avg,
             step=trainer.global_step,
             writer=trainer.vdl_writer)
-
-
-def type_name(object: object) -> str:
-    """get class name of an object"""
-    return object.__class__.__name__

ppcls/loss/__init__.py

@@ -94,6 +94,7 @@ def build_loss(config, mode="train"):
         if unlabel_loss_info:
             unlabel_train_loss_func = CombinedLoss(
                 copy.deepcopy(unlabel_loss_info))
+        return train_loss_func, unlabel_train_loss_func
     if mode == "eval" or (mode == "train" and
                           config["Global"]["eval_during_train"]):
         loss_config = config.get("Loss", None)
@@ -101,5 +102,4 @@ def build_loss(config, mode="train"):
             loss_config = loss_config.get("Eval")
             if loss_config is not None:
                 eval_loss_func = CombinedLoss(copy.deepcopy(loss_config))
-
-    return train_loss_func, unlabel_train_loss_func, eval_loss_func
+        return eval_loss_func

ppcls/metric/__init__.py

@@ -65,22 +65,19 @@ class CombinedMetrics(AvgMetrics):
                 metric.reset()
 
 
-def build_metrics(engine):
-    config, mode = engine.config, engine.mode
+def build_metrics(config, mode):
     if mode == 'train' and "Metric" in config and "Train" in config[
             "Metric"] and config["Metric"]["Train"]:
         metric_config = config["Metric"]["Train"]
-        if hasattr(engine.dataloader_dict["Train"],
-                   "collate_fn") and engine.dataloader_dict[
-                       "Train"].collate_fn is not None:
+        if config["DataLoader"]["Train"]["dataset"].get("batch_transform_ops",
+                                                        None):
             for m_idx, m in enumerate(metric_config):
                 if "TopkAcc" in m:
                     msg = f"Unable to calculate accuracy when using \"batch_transform_ops\". The metric \"{m}\" has been removed."
                     logger.warning(msg)
                     metric_config.pop(m_idx)
         train_metric_func = CombinedMetrics(copy.deepcopy(metric_config))
-    else:
-        train_metric_func = None
+        return train_metric_func
 
     if mode == "eval" or (mode == "train" and
                           config["Global"]["eval_during_train"]):
@@ -97,7 +94,4 @@ def build_metrics(engine):
             else:
                 metric_config = [{"name": "Recallk", "topk": (1, 5)}]
             eval_metric_func = CombinedMetrics(copy.deepcopy(metric_config))
-    else:
-        eval_metric_func = None
-
-    return train_metric_func, eval_metric_func
+        return eval_metric_func

ppcls/optimizer/__init__.py

@@ -21,8 +21,7 @@ import copy
 import paddle
 from typing import Dict, List
 
-from ppcls.engine.train.utils import type_name
-from ppcls.utils import logger
+from ..utils import logger, type_name
 
 from . import optimizer
 
@@ -45,14 +44,10 @@ def build_lr_scheduler(lr_config, epochs, step_each_epoch):
 
 
 # model_list is None in static graph
-def build_optimizer(engine):
-    if engine.mode != "train":
-        return None, None
-    config, max_iter, model_list = engine.config, engine.dataloader_dict[
-        "Train"].max_iter, [engine.model, engine.train_loss_func]
+def build_optimizer(config, max_iter, model_list, update_freq):
     optim_config = copy.deepcopy(config["Optimizer"])
     epochs = config["Global"]["epochs"]
-    update_freq = engine.update_freq
+    update_freq = config["Global"].get("update_freq", 1)
     step_each_epoch = max_iter // update_freq
     if isinstance(optim_config, dict):
         # convert {'name': xxx, **optim_cfg} to [{name: {scope: xxx, **optim_cfg}}]

ppcls/utils/__init__.py

@@ -26,3 +26,8 @@ from .metrics import multi_hot_encode
 from .metrics import precision_recall_fscore
 from .misc import AverageMeter
 from .save_load import init_model
+
+
+def type_name(object: object) -> str:
+    """get class name of an object"""
+    return object.__class__.__name__