2martens/masterthesis
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Changed code so that images are only saved on first batch

Browse Source 
Signed-off-by: Jim Martens <github@2martens.de>
This commit is contained in:
Jim Martens
2019-07-15 11:59:35 +02:00
parent 135144efc8
commit 13252b37b2
1 changed files with 21 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files

38
src/twomartens/masterthesis/ssd.py
View File

@ -182,11 +182,14 @@ def predict(generator: callable,
output_path=output_path, output_path=output_path,
coco_path=coco_path, coco_path=coco_path,
image_size=image_size), image_size=image_size),
decode_func=functools.partial(
_decode_predictions,
decode_func=ssd_output_decoder.decode_detections_fast,
image_size=image_size
),
transform_func=functools.partial( transform_func=functools.partial(
_transform_predictions, _transform_predictions,
decode_func=ssd_output_decoder.decode_detections_fast, inverse_transform_func=object_detection_2d_misc_utils.apply_inverse_transforms),
inverse_transform_func=object_detection_2d_misc_utils.apply_inverse_transforms,
image_size=image_size),
save_func=functools.partial(_save_predictions, save_func=functools.partial(_save_predictions,
output_file=output_file, output_file=output_file,
label_output_file=label_output_file, label_output_file=label_output_file,
@ -264,7 +267,7 @@ def _predict_prepare_paths(output_path: str, use_dropout: bool) -> Tuple[str, st
def _predict_loop(generator: Generator, use_dropout: bool, steps_per_epoch: int, def _predict_loop(generator: Generator, use_dropout: bool, steps_per_epoch: int,
dropout_step: callable, vanilla_step: callable, dropout_step: callable, vanilla_step: callable,
save_images: callable, save_images: callable, decode_func: callable,
transform_func: callable, save_func: callable) -> None: transform_func: callable, save_func: callable) -> None:
batch_counter = 0 batch_counter = 0
@ -277,11 +280,12 @@ def _predict_loop(generator: Generator, use_dropout: bool, steps_per_epoch: int,
if not saved_images: if not saved_images:
save_images(inputs, predictions, custom_string="after-prediction") save_images(inputs, predictions, custom_string="after-prediction")
decoded_predictions = decode_func(predictions)
if not saved_images:
save_images(inputs, decoded_predictions, custom_string="after-decoding")
saved_images = True saved_images = True
transformed_predictions = transform_func(predictions, transformed_predictions = transform_func(decoded_predictions,
inverse_transforms, inverse_transforms)
functools.partial(save_images,
inputs))
save_func(transformed_predictions, original_labels, filenames, save_func(transformed_predictions, original_labels, filenames,
batch_nr=batch_counter) batch_nr=batch_counter)
@ -313,21 +317,21 @@ def _predict_vanilla_step(inputs: np.ndarray, model: tf.keras.models.Model) -> n
return np.asarray(model.predict_on_batch(inputs)) return np.asarray(model.predict_on_batch(inputs))
def _transform_predictions(predictions: np.ndarray, inverse_transforms: Sequence[np.ndarray], def _decode_predictions(predictions: np.ndarray,
save_images: callable, decode_func: callable,
decode_func: callable, inverse_transform_func: callable, image_size: int) -> np.ndarray:
image_size: int) -> np.ndarray: return decode_func(
decoded_predictions = decode_func(
y_pred=predictions, y_pred=predictions,
img_width=image_size, img_width=image_size,
img_height=image_size, img_height=image_size,
input_coords="corners" input_coords="corners"
) )
save_images(decoded_predictions, custom_string="after-decoding")
transformed_predictions = inverse_transform_func(decoded_predictions, inverse_transforms)
def _transform_predictions(decoded_predictions: np.ndarray, inverse_transforms: Sequence[np.ndarray],
inverse_transform_func: callable) -> np.ndarray:
return transformed_predictions return inverse_transform_func(decoded_predictions, inverse_transforms)
def _save_predictions(transformed_predictions: np.ndarray, original_labels: np.ndarray, filenames: Sequence[str], def _save_predictions(transformed_predictions: np.ndarray, original_labels: np.ndarray, filenames: Sequence[str],