ref: organize args 2/n (#3448)

* ref: organize args 2/n

* ref: organize args 2/n

* ref: organize args 2/n

pytorch_lightning/accelerators/accelerator_connector.py

@@ -20,7 +20,8 @@ def on_trainer_init(
             num_nodes,
             log_gpu_memory,
             sync_batchnorm,
-            benchmark
+            benchmark,
+            replace_sampler_ddp
     ):
         # benchmarking
         self.trainer.benchmark = benchmark
@@ -84,6 +85,8 @@ def on_trainer_init(
 
         self.trainer.on_colab_kaggle = os.getenv('COLAB_GPU') or os.getenv('KAGGLE_URL_BASE')
 
+        self.trainer.replace_sampler_ddp = replace_sampler_ddp
+
     def select_accelerator(self):
         # SLURM ddp
         use_slurm_ddp = self.trainer.use_ddp and self.trainer.is_slurm_managing_tasks

pytorch_lightning/trainer/callback_connector.py

@@ -1,3 +1,4 @@
+import os
 from pytorch_lightning.callbacks import ModelCheckpoint, EarlyStopping, ProgressBarBase, ProgressBar
 from pytorch_lightning.utilities.exceptions import MisconfigurationException
 from pytorch_lightning.utilities.model_utils import is_overridden
@@ -15,7 +16,13 @@ def on_trainer_init(
             checkpoint_callback,
             progress_bar_refresh_rate,
             process_position,
+            default_root_dir,
+            weights_save_path
     ):
+        # init folder paths for checkpoint + weights save callbacks
+        self.trainer._default_root_dir = default_root_dir or os.getcwd()
+        self.trainer._weights_save_path = weights_save_path or self.trainer._default_root_dir
+
         # init callbacks
         self.trainer.callbacks = callbacks or []
 

pytorch_lightning/trainer/data_connector.py

@@ -24,6 +24,11 @@ class DataConnector(object):
     def __init__(self, trainer):
         self.trainer = trainer
 
+    def on_trainer_init(self, check_val_every_n_epoch, reload_dataloaders_every_epoch):
+        self.trainer.check_val_every_n_epoch = check_val_every_n_epoch
+        self.trainer.reload_dataloaders_every_epoch = reload_dataloaders_every_epoch
+        self.trainer._is_data_prepared = False
+
     def get_profiled_train_dataloader(self, train_dataloader):
         profiled_dl = self.trainer.profiler.profile_iterable(
             enumerate(self._with_is_last(train_dataloader)),

pytorch_lightning/trainer/debugging_connector.py

@@ -35,6 +35,7 @@ def on_init_start(
             overfit_batches,
             fast_dev_run
     ):
+
         self.trainer.fast_dev_run = fast_dev_run
         if self.trainer.fast_dev_run:
             limit_train_batches = 1

pytorch_lightning/trainer/logger_connector.py

@@ -13,10 +13,12 @@
 # limitations under the License.
 import torch
 from pytorch_lightning.core import memory
+from pytorch_lightning.loggers import TensorBoardLogger, LoggerCollection
 from pytorch_lightning.utilities import flatten_dict
 from pytorch_lightning.utilities.model_utils import is_overridden
 from pytorch_lightning.core.step_result import EvalResult, Result
 from pprint import pprint
+from typing import Iterable
 
 
 class LoggerConnector:
@@ -27,6 +29,28 @@ def __init__(self, trainer):
         self.logged_metrics = {}
         self.progress_bar_metrics = {}
 
+    def on_trainer_init(self, logger, log_save_interval, row_log_interval):
+        # logging
+        self.configure_logger(logger)
+        self.trainer.log_save_interval = log_save_interval
+        self.trainer.row_log_interval = row_log_interval
+
+    def configure_logger(self, logger):
+        if logger is True:
+            # default logger
+            self.trainer.logger = TensorBoardLogger(
+                save_dir=self.trainer.default_root_dir,
+                version=self.trainer.slurm_job_id,
+                name='lightning_logs'
+            )
+        elif logger is False:
+            self.trainer.logger = None
+        else:
+            if isinstance(logger, Iterable):
+                self.trainer.logger = LoggerCollection(logger)
+            else:
+                self.trainer.logger = logger
+
     def log_metrics(self, metrics, grad_norm_dic, step=None):
         """Logs the metric dict passed in.
         If `step` parameter is None and `step` key is presented is metrics,

pytorch_lightning/trainer/logging.py

@@ -39,22 +39,6 @@ class TrainerLoggingMixin(ABC):
     num_gpus: int
     logged_metrics: ...
 
-    def configure_logger(self, logger):
-        if logger is True:
-            # default logger
-            self.logger = TensorBoardLogger(
-                save_dir=self.default_root_dir,
-                version=self.slurm_job_id,
-                name='lightning_logs'
-            )
-        elif logger is False:
-            self.logger = None
-        else:
-            if isinstance(logger, Iterable):
-                self.logger = LoggerCollection(logger)
-            else:
-                self.logger = logger
-
     def metrics_to_scalars(self, metrics):
         new_metrics = {}
         for k, v in metrics.items():

pytorch_lightning/trainer/trainer.py

@@ -46,6 +46,7 @@
 from pytorch_lightning.accelerators.accelerator_connector import AcceleratorConnector
 from pytorch_lightning.trainer.logger_connector import LoggerConnector
 from pytorch_lightning.trainer.lr_scheduler_connector import LRSchedulerConnector
+from pytorch_lightning.trainer.training_trick_connector import TrainingTricksConnector
 from pytorch_lightning.trainer.callback_connector import CallbackConnector
 from pytorch_lightning.trainer.model_connector import ModelConnector
 from pytorch_lightning.trainer.debugging_connector import DebuggingConnector
@@ -177,6 +178,7 @@ def __init__(
         self.precision_connector = PrecisionConnector(self)
         self.callback_connector = CallbackConnector(self)
         self.debugging_connector = DebuggingConnector(self)
+        self.training_tricks_connector = TrainingTricksConnector(self)
 
         self.tuner = Tuner(self)
         self.accelerator_backend = None
@@ -203,6 +205,7 @@ def __init__(
         self.tested_ckpt_path = None
 
         # training state
+        self.weights_summary = weights_summary
         self.model = None
         self.datamodule = None
         self.testing = False
@@ -217,26 +220,30 @@ def __init__(
         self.running_sanity_check = False
         self._state = TrainerState.INITIALIZING
 
-        self._default_root_dir = default_root_dir or os.getcwd()
-        self._weights_save_path = weights_save_path or self._default_root_dir
-
         # init callbacks
         self.callback_connector.on_trainer_init(
             callbacks,
             early_stop_callback,
             checkpoint_callback,
             progress_bar_refresh_rate,
-            process_position
+            process_position,
+            default_root_dir,
+            weights_save_path,
         )
 
-        self.on_init_start()
+        # init data flags
+        self.data_connector.on_trainer_init(check_val_every_n_epoch, reload_dataloaders_every_epoch)
 
-        self.gradient_clip_val = gradient_clip_val
-        self.check_val_every_n_epoch = check_val_every_n_epoch
+        # hook
+        self.on_init_start()
 
-        if not isinstance(track_grad_norm, (int, float)) and track_grad_norm != 'inf':
-            raise MisconfigurationException("track_grad_norm can be an int, a float or 'inf' (infinity norm).")
-        self.track_grad_norm = float(track_grad_norm)
+        # init training tricks
+        self.training_tricks_connector.on_trainer_init(
+            gradient_clip_val,
+            track_grad_norm,
+            accumulate_grad_batches,
+            truncated_bptt_steps
+        )
 
         # init accelerator related flags
         self.accelerator_connector.on_trainer_init(
@@ -248,25 +255,16 @@ def __init__(
             num_nodes,
             log_gpu_memory,
             sync_batchnorm,
-            benchmark
+            benchmark,
+            replace_sampler_ddp
         )
 
-        # -------------------
-        # CONTINUE
-        # -------------------
-        self.weights_summary = weights_summary
-
         # init train loop related flags
         self.train_loop.on_init_start(max_epochs, min_epochs, max_steps, min_steps, num_sanity_val_steps)
 
-        self.reload_dataloaders_every_epoch = reload_dataloaders_every_epoch
-
         self.auto_lr_find = auto_lr_find
         self.auto_scale_batch_size = auto_scale_batch_size
-        self._is_data_prepared = False
-        self.replace_sampler_ddp = replace_sampler_ddp
 
-        self.truncated_bptt_steps = truncated_bptt_steps
         self.resume_from_checkpoint = resume_from_checkpoint
         self.terminate_on_nan = terminate_on_nan
         self.shown_warnings = set()
@@ -276,14 +274,8 @@ def __init__(
             profiler = SimpleProfiler()
         self.profiler = profiler or PassThroughProfiler()
 
-        # accumulated grads
-        self.accumulate_grad_batches = accumulate_grad_batches
-        self.configure_accumulated_gradients(accumulate_grad_batches)
-
-        # logging
-        self.configure_logger(logger)
-        self.log_save_interval = log_save_interval
-        self.row_log_interval = row_log_interval
+        # init logger flags
+        self.logger_connector.on_trainer_init(logger, log_save_interval, row_log_interval)
 
         # init debugging flags
         self.debugging_connector.on_init_start(

pytorch_lightning/trainer/training_trick_connector.py

@@ -0,0 +1,45 @@
+# Copyright The PyTorch Lightning team.
+#
+# 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.
+from pytorch_lightning.utilities.exceptions import MisconfigurationException
+from pytorch_lightning.callbacks import GradientAccumulationScheduler
+
+
+class TrainingTricksConnector:
+
+    def __init__(self, trainer):
+        self.trainer = trainer
+
+    def on_trainer_init(self, gradient_clip_val, track_grad_norm, accumulate_grad_batches, truncated_bptt_steps):
+        # gradient clipping
+        self.trainer.gradient_clip_val = gradient_clip_val
+
+        # gradient norm tracking
+        if not isinstance(track_grad_norm, (int, float)) and track_grad_norm != 'inf':
+            raise MisconfigurationException("track_grad_norm can be an int, a float or 'inf' (infinity norm).")
+        self.trainer.track_grad_norm = float(track_grad_norm)
+
+        # accumulated grads
+        self.trainer.accumulate_grad_batches = accumulate_grad_batches
+        self.configure_accumulated_gradients(accumulate_grad_batches)
+
+        self.trainer.truncated_bptt_steps = truncated_bptt_steps
+
+    def configure_accumulated_gradients(self, accumulate_grad_batches):
+        if isinstance(accumulate_grad_batches, dict):
+            self.trainer.accumulation_scheduler = GradientAccumulationScheduler(accumulate_grad_batches)
+        elif isinstance(accumulate_grad_batches, int):
+            schedule = {0: accumulate_grad_batches}
+            self.trainer.accumulation_scheduler = GradientAccumulationScheduler(schedule)
+        else:
+            raise TypeError("Gradient accumulation supports only int and dict types")

pytorch_lightning/trainer/training_tricks.py

@@ -76,12 +76,3 @@ def detect_nan_tensors(self, loss: Tensor) -> None:
                     f'Detected nan and/or inf values in `{name}`.'
                     ' Check your forward pass for numerically unstable operations.'
                 )
-
-    def configure_accumulated_gradients(self, accumulate_grad_batches):
-        if isinstance(accumulate_grad_batches, dict):
-            self.accumulation_scheduler = GradientAccumulationScheduler(accumulate_grad_batches)
-        elif isinstance(accumulate_grad_batches, int):
-            schedule = {0: accumulate_grad_batches}
-            self.accumulation_scheduler = GradientAccumulationScheduler(schedule)
-        else:
-            raise TypeError("Gradient accumulation supports only int and dict types")