fix(helper): add real progressbar for training (#136)

* fix(helper): add real progressbar for training

* fix(helper): add real progressbar for training

* fix(helper): add real progressbar for training

finetuner/tuner/base.py

@@ -1,4 +1,5 @@
 import abc
+import warnings
 from typing import (
     Optional,
     Union,
@@ -7,8 +8,6 @@
     Dict,
 )
 
-from jina.logging.logger import JinaLogger
-
 from ..helper import AnyDNN, AnyDataLoader, AnyOptimizer, DocumentArrayLike
 
 
@@ -46,7 +45,6 @@ def __init__(
     ):
         self._embed_model = embed_model
         self._head_layer = head_layer
-        self.logger = JinaLogger(self.__class__.__name__)
 
     def _get_optimizer_kwargs(self, optimizer: str, custom_kwargs: Optional[Dict]):
         """Merges user-provided optimizer kwargs with default ones."""
@@ -74,7 +72,7 @@ def _get_optimizer_kwargs(self, optimizer: str, custom_kwargs: Optional[Dict]):
         custom_kwargs = custom_kwargs or {}
         extra_args = set(custom_kwargs.keys()) - set(opt_kwargs.keys())
         if extra_args:
-            self.logger.warning(
+            warnings.warn(
                 f'The following arguments are not valid for the optimizer {optimizer}:'
                 f' {extra_args}'
             )

finetuner/tuner/keras/__init__.py

@@ -83,10 +83,14 @@ def _train(self, data, optimizer, description: str):
 
         log_generator = LogGenerator('T', losses, metrics)
 
+        train_data_len = 0
         with ProgressBar(
-            description, message_on_done=log_generator, final_line_feed=False
+            description,
+            message_on_done=log_generator,
+            final_line_feed=False,
+            total_length=train_data_len,
         ) as p:
-
+            train_data_len = 0
             for inputs, label in data:
                 with tf.GradientTape() as tape:
                     outputs = self.wrapped_model(inputs, training=True)
@@ -102,6 +106,7 @@ def _train(self, data, optimizer, description: str):
                 metrics.append(metric.numpy())
 
                 p.update(message=log_generator())
+                train_data_len += 1
 
         return losses, metrics
 

finetuner/tuner/paddle/__init__.py

@@ -99,10 +99,14 @@ def _train(self, data, optimizer: Optimizer, description: str):
         metrics = []
 
         log_generator = LogGenerator('T', losses, metrics)
-
+        train_data_len = 0
         with ProgressBar(
-            description, message_on_done=log_generator, final_line_feed=False
+            description,
+            message_on_done=log_generator,
+            final_line_feed=False,
+            total_length=train_data_len,
         ) as p:
+            train_data_len = 0
             for inputs, label in data:
                 # forward step
                 outputs = self.wrapped_model(*inputs)
@@ -118,6 +122,7 @@ def _train(self, data, optimizer: Optimizer, description: str):
                 metrics.append(metric.numpy())
 
                 p.update(message=log_generator())
+                train_data_len += 1
         return losses, metrics
 
     def fit(

finetuner/tuner/pytorch/__init__.py

@@ -106,10 +106,14 @@ def _train(self, data, optimizer: Optimizer, description: str):
         metrics = []
 
         log_generator = LogGenerator('T', losses, metrics)
-
+        train_data_len = 0
         with ProgressBar(
-            description, message_on_done=log_generator, final_line_feed=False
+            description,
+            message_on_done=log_generator,
+            final_line_feed=False,
+            total_length=train_data_len,
         ) as p:
+            train_data_len = 0
             for inputs, label in data:
                 # forward step
                 inputs = [inpt.to(self.device) for inpt in inputs]
@@ -128,6 +132,7 @@ def _train(self, data, optimizer: Optimizer, description: str):
                 metrics.append(metric.cpu().numpy())
 
                 p.update(message=log_generator())
+                train_data_len += 1
         return losses, metrics
 
     def fit(