def construct_input_fns(dataset, data_dir, batch_size=16, repeat=1):
"""Construct train and test input functions, as well as the column fn."""
if _check_buffers(data_dir, dataset):
train_df, eval_df = None, None
else:
df = movielens.csv_to_joint_dataframe(dataset=dataset, data_dir=data_dir)
df = movielens.integerize_genres(dataframe=df)
df = df.drop(columns=[movielens.TITLE_COLUMN])
train_df = df.sample(frac=0.8, random_state=0)
eval_df = df.drop(train_df.index)
train_df = train_df.reset_index(drop=True)
eval_df = eval_df.reset_index(drop=True)
train_input_fn = _df_to_input_fn(
df=train_df, name="train", dataset=dataset, data_dir=data_dir,
batch_size=batch_size, repeat=repeat,
shuffle=movielens.NUM_RATINGS[dataset])
eval_input_fn = _df_to_input_fn(
df=eval_df, name="eval", dataset=dataset, data_dir=data_dir,
batch_size=batch_size, repeat=repeat, shuffle=None)
model_column_fn = functools.partial(build_model_columns, dataset=dataset)
train_input_fn()
return train_input_fn, eval_input_fn, model_column_fn
def construct_input_fns(dataset, data_dir, batch_size=16, repeat=1):
"""Construct train and test input functions, as well as the column fn."""
if _check_buffers(data_dir, dataset):
train_df, eval_df = None, None
else:
df = movielens.csv_to_joint_dataframe(dataset=dataset,
data_dir=data_dir)
df = movielens.integerize_genres(dataframe=df)
df = df.drop(columns=[movielens.TITLE_COLUMN])
train_df = df.sample(frac=0.8, random_state=0)
eval_df = df.drop(train_df.index)
train_df = train_df.reset_index(drop=True)
eval_df = eval_df.reset_index(drop=True)
train_input_fn = _df_to_input_fn(df=train_df,
name="train",
dataset=dataset,
data_dir=data_dir,
batch_size=batch_size,
repeat=repeat,
shuffle=movielens.NUM_RATINGS[dataset])
eval_input_fn = _df_to_input_fn(df=eval_df,
name="eval",
dataset=dataset,
data_dir=data_dir,
batch_size=batch_size,
repeat=repeat,
shuffle=None)
model_column_fn = functools.partial(build_model_columns, dataset=dataset)
train_input_fn()
return train_input_fn, eval_input_fn, model_column_fn