Datasets:
celinelee
/
python_fn_calls

Dataset card Data Studio Files
xet
 Community
1
code
stringlengths
3
6.57k
learning_rate.assign(tf.maximum(learning_rate_decay * learning_rate, min_learning_rate)
ValueError('Unknown optimizer: {}'.format(optimizer)
tf_optimizer_class(learning_rate=learning_rate)
os.path.exists(tensorboard_folder)
shutil.rmtree(tensorboard_folder)
tf.summary.FileWriter(tensorboard_folder)
LossHook(reader, iter_interval, summary_writer=sw)
ETAHook(reader, iter_interval, int(math.ceil(len(train_data)
ExamplesPerSecHook(reader, batch_size, iter_interval, sw)
preferred_metric_and_initial_score()
side_effect(metrics, prev_metric)
state (in this case a metric)
reader.store(save_dir)
reader.session.run(lr_decay_op)
logger.info("Decayed learning rate to: %.5f" % reader.session.run(learning_rate)
reader.model_module.store(os.path.join(save_dir, "model_module")
logger.info("Saving reader to: %s" % save_dir)
reader.load(save_dir)
evaluate_reader(reader, dev_data, batch_size)
logger.info("############### Results on the Dev Set##############")
pretty_print_results(result_dict)
reader.load(save_dir)
evaluate_reader(reader, test_data, batch_size)
logger.info("############### Results on the Test Set##############")
pretty_print_results(result_dict)
train_pytorch(reader, train_data, test_data, dev_data, configuration: dict, debug=False)
configuration.get('seed')
random.seed(seed)
torch.manual_seed(seed)
configuration.get('clip_value')
configuration.get('batch_size')
configuration.get('epochs')
configuration.get('l2')
configuration.get('optimizer')
configuration.get('learning_rate')
configuration.get('learning_rate_decay')
configuration.get('log_interval')
configuration.get('validation_interval')
configuration.get('tensorboard_folder')
configuration.get('reader')
configuration.get('save_dir')
configuration.get('write_metrics_to')
reader.setup_from_data(train_data, is_training=True)
abs(clip_value)
abs(clip_value)
ValueError('Unknown optimizer: {}'.format(optimizer)
list(reader.model_module.prediction_module.parameters()
params.extend(reader.model_module.loss_module.parameters()
torch_optimizer_class(params, lr=learning_rate)
os.path.exists(tensorboard_folder)
shutil.rmtree(tensorboard_folder)
tf.summary.FileWriter(tensorboard_folder)
LossHook(reader, iter_interval, summary_writer=sw)
ExamplesPerSecHook(reader, batch_size, iter_interval, sw)
preferred_metric_and_initial_score()
side_effect(metrics, prev_metric)
state (in this case a metric)
logger.info("Decayed learning rate to: %.5f" % param_group['lr'])
reader.store(save_dir)
reader.model_module.store(os.path.join(save_dir, "model_module")
logger.info("Saving model to: %s" % save_dir)
reader.load(save_dir)
evaluate_reader(reader, dev_data, batch_size)
logger.info("############### Results on the Dev Set##############")
pretty_print_results(result_dict)
reader.load(save_dir)
evaluate_reader(reader, test_data, batch_size)
logger.info("############### Results on the Test Set##############")
pretty_print_results(result_dict)
BatchNormLayer(nn.Module)
parameters (e.g., the outputs of a hypernetwork)
alone (i.e., no weights being passed to the :meth:`forward` method)
parameters (:code:`requires_grad=True`)
super(BatchNormLayer, self)
__init__()
self.register_buffer('_num_stats', torch.tensor(0, dtype=torch.long)
nn.ParameterList()
torch.Tensor(num_features)
torch.Tensor(num_features)
self._weights.append(self.scale)
self._weights.append(self.bias)
nn.init.ones_(self.scale)
nn.init.zeros_(self.bias)
NotImplementedError()
self.checkpoint_stats()
self._stats_names(0)
self.register_buffer(mname, None)
self.register_buffer(vname, None)
weights(self)
param_shapes(self)
hyper_shapes(self)
NotImplementedError('Not implemented yet!')
num_stats(self)
module (attribute :attr:`training`)
used (denoted by *use batch stats* in the table below)
training (denoted by *update running stats* in the table below)
batch (also in eval mode)
running_mean (optional)
running_var (optional)
weight (optional)