drop

Lightning-AI · Dec 8, 2020 · 99ff309 · 99ff309
1 parent aeaa6b2
commit 99ff309
Show file tree

Hide file tree

Showing 3 changed files with 5 additions and 120 deletions.
diff --git a/pytorch_lightning/trainer/connectors/logger_connector/logger_connector.py b/pytorch_lightning/trainer/connectors/logger_connector/logger_connector.py
@@ -392,31 +392,14 @@ def log_train_epoch_end_metrics(
         # [optimizer_idx][training_step_idx][tbptt_index]
         opt_idx_outputs = epoch_output[0]
 
-        # TODO: deprecate 1.0
-        try:
-            sample_obj = opt_idx_outputs[0][0] if isinstance(opt_idx_outputs[0], list) else opt_idx_outputs[0]
-            is_result_obj = len(epoch_output) > 0 and isinstance(sample_obj, Result)
-            is_1_0_result = is_result_obj and 'extra' in sample_obj
-        except IndexError as e:
-            is_result_obj = False
-            is_1_0_result = False
-
         # ------------------
         # NEW 1.0.0 PATH
         # ------------------
-        if is_1_0_result:
-            # lightning module hook
-            self.training_epoch_end(model, epoch_output, num_optimizers)
-
-            # log/aggregate metrics automatically
-            epoch_log_metrics, epoch_progress_bar_metrics = self.__auto_reduce_results_on_epoch_end(epoch_output)
+        # lightning module hook
+        self.training_epoch_end(model, epoch_output, num_optimizers)
 
-        # TODO: deprecate 1.0
-        else:
-            out = self.__run_legacy_training_epoch_end(
-                num_optimizers, epoch_output, model, is_result_obj, epoch_callback_metrics
-            )
-            epoch_log_metrics, epoch_progress_bar_metrics, epoch_callback_metrics = out
+        # log/aggregate metrics automatically
+        epoch_log_metrics, epoch_progress_bar_metrics = self.__auto_reduce_results_on_epoch_end(epoch_output)
 
         # it will perform reduction over epoch and return log metrics
         cached_epoch_log_metrics = self.cached_results.get_epoch_log_metrics()

diff --git a/pytorch_lightning/trainer/logging.py b/pytorch_lightning/trainer/logging.py
@@ -59,23 +59,6 @@ def process_dict_result(self, output, train=False):
 
         Separates loss from logging and progress bar metrics
         """
-        # --------------------
-        # WARN DEPRECATED KEYS
-        # --------------------
-        # TODO: 1.0.0 remove
-        if isinstance(output, dict):
-            for k, v in output.items():
-                if k in ['log', 'progress_bar']:
-                    m = inspect.cleandoc(
-                        f"""The {{{k}:dict keyword}} was deprecated in 0.9.1 and will be removed in 1.0.0
-                        Please use self.log(...) inside the lightningModule instead.
-
-                        # log on a step or aggregate epoch metric to the logger and/or progress bar
-                        # (inside LightningModule)
-                        self.log('train_loss', loss, on_step=True, on_epoch=True, prog_bar=True)
-                    """)
-                    rank_zero_warn(m)
-
         # --------------------------
         # handle single scalar only
         # --------------------------

diff --git a/pytorch_lightning/trainer/training_loop.py b/pytorch_lightning/trainer/training_loop.py
@@ -373,24 +373,7 @@ def _process_training_step_output(self, training_step_output, split_batch):
             return None, None
 
         # -----------------------------------------
-        # process result return (DEPRECATE in 1.0)
-        # -----------------------------------------
-        if isinstance(training_step_output, Result):
-            training_step_output_for_epoch_end = self._process_result(training_step_output, split_batch)
-            return training_step_output_for_epoch_end, training_step_output
-
-        # -----------------------------------------
-        # process hybrid (1.0)
-        # -----------------------------------------
-        # no need for these checks in 1.0.0
-        # TODO: remove checks in 1.0.0
-        is_tensor = isinstance(training_step_output_for_epoch_end, torch.Tensor)
-        is_1_0_output = is_tensor or ("log" not in training_step_output and "progress_bar" not in training_step_output)
-        if is_1_0_output:
-            return self._process_training_step_output_1_0(training_step_output, split_batch)
-
-        # -----------------------------------------
-        # process old dict (deprecate 1.0)
+        # process old dict
         # -----------------------------------------
         training_step_output = self.trainer.process_dict_result(training_step_output, train=True)
 
@@ -409,70 +392,6 @@ def _process_training_step_output(self, training_step_output, split_batch):
 
         return training_step_output_for_epoch_end, training_step_output
 
-    def _process_training_step_output_1_0(self, training_step_output, split_batch):
-        result = self.trainer.get_model()._results
-
-        loss = None
-        hiddens = None
-
-        # handle dict return
-        if isinstance(training_step_output, dict):
-            loss = training_step_output.pop("loss", None)
-            hiddens = training_step_output.pop("hiddens", None)
-            result["extra"] = training_step_output
-
-        # handle scalar return
-        elif isinstance(training_step_output, torch.Tensor):
-            loss = training_step_output
-            result["extra"] = {}
-
-        # map to results under the hood
-        result.minimize = loss
-        result.hiddens = hiddens
-
-        # track batch for manual reduction with result
-        result.track_batch_size(len(split_batch))
-
-        # track metrics without grads for epoch reduction
-        training_step_output_for_epoch_end = copy(result)
-        training_step_output_for_epoch_end.detach()
-        if self.trainer.move_metrics_to_cpu:
-            training_step_output_for_epoch_end.cpu()
-
-        # what flows back into the system
-        training_step_output = result
-
-        return training_step_output_for_epoch_end, training_step_output
-
-    def _process_result(self, training_step_output, split_batch):
-        training_step_output.track_batch_size(len(split_batch))
-        m = """
-        TrainResult and EvalResult were deprecated in 0.9.1 and support will drop in 1.0.0.
-        Use self.log and .write from the LightningModule to log metrics and write predictions.
-        training_step can now only return a scalar (for the loss) or a dictionary with anything you want.
-
-        Option 1:
-        return loss
-
-        Option 2:
-        return {'loss': loss, 'anything_else': ...}
-
-        Option 3:
-        return {'loss': loss, 'hiddens': hiddens, 'anything_else': ...}
-            """
-        rank_zero_warn(m)
-
-        # don't allow EvalResult in the training_step
-        if isinstance(training_step_output, EvalResult):
-            raise MisconfigurationException(
-                "training_step cannot return EvalResult, " "use a dict or TrainResult instead"
-            )
-
-        training_step_output_for_epoch_end = copy(training_step_output)
-        training_step_output_for_epoch_end.detach()
-
-        return training_step_output_for_epoch_end
-
     def optimizer_step(self, optimizer, opt_idx, batch_idx, train_step_and_backward_closure, *args, **kwargs):
         model_ref = self.trainer.get_model()