Fixed pylint error (intel#18)

* Fixed pylint error

examples/optimize/pytorch/huggingface/question-answering/pruning/requirements.txt

@@ -1,4 +1,4 @@
 datasets >= 1.8.0
 torch >= 1.10.0
 transformers>=4.12.0
-wandb
+wandb

nlp_toolkit/optimization/auto_distillation.py

@@ -145,10 +145,10 @@ def train_evaluate(self, model):
         Returns:
             Evaluated metrics of the model.
         """
-        assert self.train_func is not None and self.eval_func is not None, \
+        assert self._train_func is not None and self._eval_func is not None, \
             "train_func and eval_func must be set."
-        model = self.train_func(model)
-        return self.eval_func(model)
+        model = self._train_func(model)
+        return self._eval_func(model)
 
     def metrics_conversion(self, metrics):
         if isinstance(metrics, dict):

nlp_toolkit/optimization/config.py

@@ -491,7 +491,7 @@ def __init__(
         metrics: Union[List, Metric] = None,
     ):
         super().__init__()
-        from nncf import NNCFConfig
+        from nncf import NNCFConfig # disable=E0401
         assert isinstance(nncf_config, NNCFConfig)
         self.nncf_config = nncf_config
         if metrics is not None:

nlp_toolkit/optimization/model.py

@@ -66,7 +66,10 @@ def from_pretrained(
             ("resume_download", False),
             ("revision", None),
         ]
-        download_kwargs = {name: kwargs.get(name, default_value) for (name, default_value) in download_kwarg_default}
+        download_kwargs = {
+            name: kwargs.get(name, default_value)
+            for (name, default_value) in download_kwarg_default
+        }
 
         config = AutoConfig.from_pretrained(model_name_or_path)
         model_class = eval(f'transformers.{config.architectures[0]}')
@@ -82,7 +85,8 @@ def from_pretrained(
             keys_to_ignore_on_load_unexpected = copy.deepcopy(
                 getattr(model_class, "_keys_to_ignore_on_load_unexpected", None)
             )
-            keys_to_ignore_on_load_missing = copy.deepcopy(getattr(model_class, "_keys_to_ignore_on_load_missing", None))
+            keys_to_ignore_on_load_missing = \
+                copy.deepcopy(getattr(model_class, "_keys_to_ignore_on_load_missing", None))
 
             # Avoid unnecessary warnings resulting from quantized model initialization
             quantized_keys_to_ignore_on_load = [r"zero_point", r"scale", 
@@ -91,7 +95,9 @@ def from_pretrained(
             if keys_to_ignore_on_load_unexpected is None:
                 model_class._keys_to_ignore_on_load_unexpected = quantized_keys_to_ignore_on_load
             else:
-                model_class._keys_to_ignore_on_load_unexpected.extend(quantized_keys_to_ignore_on_load)
+                model_class._keys_to_ignore_on_load_unexpected.extend(
+                    quantized_keys_to_ignore_on_load
+                )
             missing_keys_to_ignore_on_load = [r"weight", r"bias"]
             if keys_to_ignore_on_load_missing is None:
                 model_class._keys_to_ignore_on_load_missing = missing_keys_to_ignore_on_load
@@ -104,7 +110,9 @@ def from_pretrained(
             model_class._keys_to_ignore_on_load_missing = keys_to_ignore_on_load_missing
 
             if not os.path.isdir(model_name_or_path) and not os.path.isfile(model_name_or_path):
-                config_file = hf_bucket_url(model_name_or_path, filename="best_configure.yaml", revision=download_kwargs["revision]"])
+                config_file = hf_bucket_url(model_name_or_path,
+                                            filename="best_configure.yaml",
+                                            revision=download_kwargs["revision]"])
 
                 try:
                     resolved_config_file = cached_path(
@@ -117,19 +125,25 @@ def from_pretrained(
                     logger.error(err)
                     msg = (
                         f"Can't load config for '{model_name_or_path}'. Make sure that:\n\n"
-                        f"-'{model_name_or_path}' is a correct model identifier listed on 'https://huggingface.co/models'\n\n"
-                        f"-or '{model_name_or_path}' is a correct path to a directory containing a best_configure.yaml file\n\n"
+                        f"-'{model_name_or_path}' is a correct model identifier listed on "
+                        f"'https://huggingface.co/models'\n\n"
+                        f"-or '{model_name_or_path}' is a correct path to a directory containing "
+                        f"a best_configure.yaml file\n\n"
                     )
 
                     if download_kwargs["revision]"] is not None:
                         msg += (
-                            f"- or {download_kwargs['revision']} is a valid git identifier (branch name, a tag name, or a commit id) that "
-                            f"exists for this model name as listed on its model page on 'https://huggingface.co/models'\n\n"
+                            f"- or {download_kwargs['revision']} is a valid git "
+                            f"identifier (branch name, a tag name, or a commit id) that "
+                            f"exists for this model name as listed on its model page on "
+                            f"'https://huggingface.co/models'\n\n"
                         )
 
                     raise EnvironmentError(msg)
 
-                config_file = hf_bucket_url(model_name_or_path, filename="best_model_weights.pt", revision=download_kwargs["revision]"])
+                config_file = hf_bucket_url(model_name_or_path,
+                                            filename="best_model_weights.pt",
+                                            revision=download_kwargs["revision]"])
                 try:
                     resolved_config_file = cached_path(
                         config_file,
@@ -141,14 +155,18 @@ def from_pretrained(
                     logger.error(err)
                     msg = (
                         f"Can't load config for '{model_name_or_path}'. Make sure that:\n\n"
-                        f"-'{model_name_or_path}' is a correct model identifier listed on 'https://huggingface.co/models'\n\n"
-                        f"-or '{model_name_or_path}' is a correct path to a directory containing a best_model_weights.pt file\n\n"
+                        f"-'{model_name_or_path}' is a correct model identifier listed on "
+                        f"'https://huggingface.co/models'\n\n"
+                        f"-or '{model_name_or_path}' is a correct path to a directory containing "
+                        f"a best_model_weights.pt file\n\n"
                     )
 
                     if download_kwargs["revision]"] is not None:
                         msg += (
-                            f"- or {download_kwargs['revision']} is a valid git identifier (branch name, a tag name, or a commit id) that "
-                            f"exists for this model name as listed on its model page on 'https://huggingface.co/models'\n\n"
+                            f"- or {download_kwargs['revision']} is a valid git identifier "
+                            f"(branch name, a tag name, or a commit id) that "
+                            f"exists for this model name as listed on its model page on "
+                            f"'https://huggingface.co/models'\n\n"
                         )
 
                     raise EnvironmentError(msg)

nlp_toolkit/optimization/optimizer.py

@@ -13,6 +13,7 @@
 #  limitations under the License.
 
 import logging
+import os
 import torch
 
 from neural_compressor.experimental import common, Component, Distillation
@@ -100,6 +101,10 @@ def __init__(
         self._train_func = None
         self._calib_dataloader = None
         self.output_dir = output_dir
+        self.quant_config = None
+        self.pruning_config = None
+        self.distillation_config = None
+        self._provider = Provider.INC.value
 
     @property
     def eval_func(self):
@@ -125,7 +130,7 @@ def train_func(self, func: Callable):
     def calib_dataloader(self, dataloader):
         self._calib_dataloader = dataloader
 
-    def _init_quantize(self):
+    def _init_quantizer(self):
         from .quantization import QuantizationMode
         from neural_compressor.experimental import Quantization, common
         assert isinstance(self.quant_config, QuantizationConfig), \
@@ -157,7 +162,7 @@ def _init_quantize(self):
         self.quantizer = quantizer
         return quantizer
 
-    def _nncf_quantize(self):
+    def _nncf_quantize(self): # disable=E0401
         from nlp_toolkit import NncfConfig
         from nncf import create_compressed_model
         assert isinstance(self.quant_config, NncfConfig), \
@@ -175,7 +180,8 @@ def _nncf_quantize(self):
         )
 
         self.compression_ctrl = \
-            compression_algo_controller.distributed() if self.quant_config.distributed else compression_algo_controller
+            compression_algo_controller.distributed() if self.quant_config.distributed \
+            else compression_algo_controller
 
     def quantize(
         self,
@@ -197,9 +203,9 @@ def quantize(
             self._calib_dataloader = calib_dataloader
         quantizer = self._init_quantizer()
         opt_model = quantizer.fit()
-        opt_model.save(self.args.output_dir)
+        opt_model.save(self.output_dir)
         logger.info(
-            "quantized model and configure file have saved to {}".format(self.args.output_dir)
+            "quantized model and configure file have saved to {}".format(self.output_dir)
         )
         return opt_model.model
 
@@ -209,20 +215,6 @@ def _init_pruner(self):
         assert isinstance(self.pruning_config, PruningConfig), \
             "please pass a instance of PruningConfig to NoTrainerOptimizer.prune!"
 
-        pruning_start_epoch, pruning_end_epoch = self.pruning_config.epoch_range
-
-        if pruning_start_epoch > self.args.num_train_epochs - 1:
-            logger.warning(
-                f"Pruning end epoch {pruning_start_epoch} is higher than the total number of training epoch "
-                f"{self.args.num_train_epochs}. No pruning will be applied."
-            )
-
-        if pruning_end_epoch > self.args.num_train_epochs - 1:
-            logger.warning(
-                f"Pruning end epoch {pruning_end_epoch} is higher than the total number of training epoch "
-                f"{self.args.num_train_epochs}. The target sparsity will not be reached."
-            )
-
         pruner = Pruning(self.pruning_config.inc_config)
         pruner.model = common.Model(self.model)
 

nlp_toolkit/optimization/trainer.py

@@ -61,13 +61,13 @@
     import datasets
 
 if is_torch_tpu_available():
-    import torch_xla.core.xla_model as xm
+    import torch_xla.core.xla_model as xm # disable=E0401
 
 if is_apex_available():
-    from apex import amp
+    from apex import amp # disable=E0401
 
 if is_sagemaker_mp_enabled():
-    from .trainer_pt_utils import smp_forward_backward
+    from .trainer_pt_utils import smp_forward_backward # disable=E0401
 
 if TYPE_CHECKING:
     import optuna
@@ -189,7 +189,7 @@ def _init_quantizer(self):
         self.quantizer = quantizer
         return quantizer
 
-    def _nncf_quantize(self):
+    def _nncf_quantize(self): # disable=E0401
         from nlp_toolkit import NncfConfig
         from nncf import create_compressed_model
         compression_state = None
@@ -272,13 +272,15 @@ def _init_pruner(self):
 
         if pruning_start_epoch > self.args.num_train_epochs - 1:
             logger.warning(
-                f"Pruning end epoch {pruning_start_epoch} is higher than the total number of training epoch "
+                f"Pruning end epoch {pruning_start_epoch} is higher than "
+                f"the total number of training epoch "
                 f"{self.args.num_train_epochs}. No pruning will be applied."
             )
 
         if pruning_end_epoch > self.args.num_train_epochs - 1:
             logger.warning(
-                f"Pruning end epoch {pruning_end_epoch} is higher than the total number of training epoch "
+                f"Pruning end epoch {pruning_end_epoch} is higher than "
+                f"the total number of training epoch "
                 f"{self.args.num_train_epochs}. The target sparsity will not be reached."
             )
 
@@ -331,7 +333,8 @@ def _init_distiller(self):
         if self._eval_func is not None:
             distiller.eval_func = self._eval_func
         else:
-            assert self.metrics is not None, "Please pass metrics to trainer.distillation.metrics!"
+            assert self.metrics is not None, \
+                "Please pass metrics to trainer.distillation.metrics!"
             distiller.eval_func = self.builtin_eval_func
 
         distiller.train_func = \
@@ -861,7 +864,8 @@ def training_step(self, model: nn.Module, inputs: Dict[str, Union[torch.Tensor,
             loss_mb = smp_forward_backward(model, inputs, self.args.gradient_accumulation_steps, scaler=scaler)
             return loss_mb.reduce_mean().detach().to(self.args.device)
 
-        if self.use_amp:
+        if self.use_amp: # disable=E0401
+            from torch.cuda.amp import autocast
             with autocast():
                 loss = self.compute_loss(model, inputs)
         else:

nlp_toolkit/preprocessing/data_augmentation.py

@@ -232,8 +232,8 @@ def text_generation_augmentation(self, extension, raw_datasets):
         min_length = m - std
 
         total_count = sum(label2count.values())
-        factor = total_count \
-            if num_return_sentences <= 0 else int(math.ceil(num_return_sentences / num_samples))
+        factor = total_count if num_return_sentences <= 0 \
+                 else int(math.ceil(num_return_sentences / self._num_samples))
         p0 = label2count[0] / total_count
         p1 = 1 - p0