-
Notifications
You must be signed in to change notification settings - Fork 7
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
Showing
4 changed files
with
512 additions
and
0 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 1,251 @@ | ||
| """ | ||
| Augmentations for SimPer (and other SSL methods). | ||
| """ | ||
| import tensorflow as tf | ||
| import tensorflow_probability as tfp | ||
| from typing import Optional, Tuple | ||
|
|
||
|
|
||
| def random_apply(func, p, x): | ||
| return tf.cond( | ||
| tf.less(tf.random.uniform([], minval=0, maxval=1, dtype=tf.float32), | ||
| tf.cast(p, tf.float32)), | ||
| lambda: func(x), | ||
| lambda: x) | ||
|
|
||
|
|
||
| def resize_and_rescale(x, y, img_size): | ||
| x = tf.cast(x, tf.float32) | ||
| x = tf.image.resize(x, [img_size, img_size]) | ||
| return x, y | ||
|
|
||
|
|
||
| def _sample_or_pad_sequence_indices(sequence: tf.Tensor, num_steps: int, | ||
| stride: int, | ||
| offset: tf.Tensor) -> tf.Tensor: | ||
| sequence_length = tf.shape(sequence)[0] | ||
| sel_idx = tf.range(sequence_length) | ||
|
|
||
| max_length = num_steps * stride offset | ||
| num_repeats = tf.math.floordiv(max_length sequence_length - 1, | ||
| sequence_length) | ||
| sel_idx = tf.tile(sel_idx, [num_repeats]) | ||
|
|
||
| steps = tf.range(offset, offset num_steps * stride, stride) | ||
| return tf.gather(sel_idx, steps) | ||
|
|
||
|
|
||
| def sample_sequence(sequence: tf.Tensor, | ||
| num_steps: int, | ||
| random: bool = True, | ||
| stride: int = 1, | ||
| seed: Optional[int] = None) -> tf.Tensor: | ||
| sequence_length = tf.shape(sequence)[0] | ||
|
|
||
| if random: | ||
| sequence_length = tf.cast(sequence_length, tf.float32) | ||
| frame_stride = tf.cast(stride, tf.float32) | ||
| max_offset = tf.cond( | ||
| sequence_length > (num_steps - 1) * frame_stride, | ||
| lambda: sequence_length - (num_steps - 1) * frame_stride, | ||
| lambda: sequence_length) | ||
| offset = tf.random.uniform((), | ||
| maxval=tf.cast(max_offset, dtype=tf.int32), | ||
| dtype=tf.int32, | ||
| seed=seed) | ||
| else: | ||
| raise NotImplementedError(f"Only `random == True` is supported now.") | ||
| offset = (sequence_length - num_steps * stride) // 2 | ||
| offset = tf.maximum(0, offset) | ||
|
|
||
| indices = _sample_or_pad_sequence_indices( | ||
| sequence=sequence, num_steps=num_steps, stride=stride, offset=offset) | ||
| indices.set_shape((num_steps,)) | ||
|
|
||
| return tf.gather(sequence, indices) | ||
|
|
||
|
|
||
| def random_crop_resize(frames: tf.Tensor, | ||
| output_h: int, | ||
| output_w: int, | ||
| aspect_ratio: Tuple[float, float] = (0.75, 1.33), | ||
| area_range: Tuple[float, float] = (0.5, 1)) -> tf.Tensor: | ||
| shape = tf.shape(frames) | ||
| seq_len, _, _, channels = shape[0], shape[1], shape[2], shape[3] | ||
| bbox = tf.constant([0.0, 0.0, 1.0, 1.0], dtype=tf.float32, shape=[1, 1, 4]) | ||
| factor = output_w / output_h | ||
| aspect_ratio = (aspect_ratio[0] * factor, aspect_ratio[1] * factor) | ||
| sample_distorted_bbox = tf.image.sample_distorted_bounding_box( | ||
| shape[1:], | ||
| bounding_boxes=bbox, | ||
| min_object_covered=0.1, | ||
| aspect_ratio_range=aspect_ratio, | ||
| area_range=area_range, | ||
| max_attempts=100, | ||
| use_image_if_no_bounding_boxes=True) | ||
| bbox_begin, bbox_size, _ = sample_distorted_bbox | ||
| offset_y, offset_x, _ = tf.unstack(bbox_begin) | ||
| target_height, target_width, _ = tf.unstack(bbox_size) | ||
| size = tf.convert_to_tensor((seq_len, target_height, target_width, channels)) | ||
| offset = tf.convert_to_tensor((0, offset_y, offset_x, 0)) | ||
| frames = tf.slice(frames, offset, size) | ||
| frames = tf.cast(tf.image.resize(frames, (output_h, output_w)), frames.dtype) | ||
| return frames | ||
|
|
||
|
|
||
| def gaussian_blur(image, kernel_size, sigma, padding='SAME'): | ||
| radius = tf.cast(kernel_size / 2, dtype=tf.int32) | ||
| kernel_size = radius * 2 1 | ||
| x = tf.cast(tf.range(-radius, radius 1), dtype=tf.float32) | ||
| blur_filter = tf.exp( | ||
| -tf.pow(x, 2.0) / (2.0 * tf.pow(tf.cast(sigma, dtype=tf.float32), 2.0))) | ||
| blur_filter /= tf.reduce_sum(blur_filter) | ||
| # One vertical and one horizontal filter. | ||
| blur_v = tf.reshape(blur_filter, [kernel_size, 1, 1, 1]) | ||
| blur_h = tf.reshape(blur_filter, [1, kernel_size, 1, 1]) | ||
| num_channels = tf.shape(image)[-1] | ||
| blur_h = tf.tile(blur_h, [1, 1, num_channels, 1]) | ||
| blur_v = tf.tile(blur_v, [1, 1, num_channels, 1]) | ||
| expand_batch_dim = image.shape.ndims == 3 | ||
| if expand_batch_dim: | ||
| image = tf.expand_dims(image, axis=0) | ||
| blurred = tf.nn.depthwise_conv2d( | ||
| image, blur_h, strides=[1, 1, 1, 1], padding=padding) | ||
| blurred = tf.nn.depthwise_conv2d( | ||
| blurred, blur_v, strides=[1, 1, 1, 1], padding=padding) | ||
| if expand_batch_dim: | ||
| blurred = tf.squeeze(blurred, axis=0) | ||
| return blurred | ||
|
|
||
|
|
||
| def random_blur(image, height, p=0.2): | ||
| def _transform(image): | ||
| sigma = tf.random.uniform([], 0.1, 2.0, dtype=tf.float32) | ||
| return gaussian_blur( | ||
| image, kernel_size=height // 20, sigma=sigma, padding='SAME') | ||
|
|
||
| return random_apply(_transform, p=p, x=image) | ||
|
|
||
|
|
||
| def random_flip_left_right(frames: tf.Tensor, seed: Optional[int] = None) -> tf.Tensor: | ||
| is_flipped = tf.random.uniform((), minval=0, maxval=2, dtype=tf.int32, seed=seed) | ||
| frames = tf.cond( | ||
| tf.equal(is_flipped, 1), | ||
| true_fn=lambda: tf.image.flip_left_right(frames), | ||
| false_fn=lambda: frames) | ||
| return frames | ||
|
|
||
|
|
||
| def random_flip_up_down(frames: tf.Tensor, seed: Optional[int] = None) -> tf.Tensor: | ||
| is_flipped = tf.random.uniform((), minval=0, maxval=2, dtype=tf.int32, seed=seed) | ||
| frames = tf.cond( | ||
| tf.equal(is_flipped, 1), | ||
| true_fn=lambda: tf.image.flip_up_down(frames), | ||
| false_fn=lambda: frames) | ||
| return frames | ||
|
|
||
|
|
||
| def random_rotation(frames: tf.Tensor, seed: Optional[int] = None) -> tf.Tensor: | ||
| is_flipped = tf.random.uniform((), minval=0, maxval=2, dtype=tf.int32, seed=seed) | ||
| frames = tf.cond( | ||
| tf.equal(is_flipped, 1), | ||
| true_fn=lambda: tf.image.rot90(frames), | ||
| false_fn=lambda: frames) | ||
| return frames | ||
|
|
||
|
|
||
| def to_grayscale(image, keep_channels=True): | ||
| image = tf.image.rgb_to_grayscale(image) | ||
| if keep_channels: | ||
| image = tf.tile(image, [1, 1, 3]) | ||
| return image | ||
|
|
||
|
|
||
| def random_grayscale_3d(frames, p=0.2): | ||
| num_frames, width, height, channels = frames.shape.as_list() | ||
| big_image = tf.reshape(frames, [num_frames * width, height, channels]) | ||
| big_image = random_apply(to_grayscale, p=p, x=big_image) | ||
| return tf.reshape(big_image, [num_frames, width, height, channels]) | ||
|
|
||
|
|
||
| def random_brightness(image, max_delta=0.3): | ||
| factor = tf.random.uniform( | ||
| [], tf.maximum(1.0 - max_delta, 0), 1.0 max_delta) | ||
| image = image * factor | ||
| return image | ||
|
|
||
|
|
||
| def random_reverse(frames: tf.Tensor, seed: Optional[int] = None) -> tf.Tensor: | ||
| is_flipped = tf.random.uniform((), minval=0, maxval=2, dtype=tf.int32, seed=seed) | ||
| frames = tf.cond( | ||
| tf.equal(is_flipped, 1), | ||
| true_fn=lambda: tf.experimental.numpy.flip(frames, axis=0), | ||
| false_fn=lambda: frames) | ||
| return frames | ||
|
|
||
|
|
||
| # Arbitrary speed / frequency augmentation for SimPer | ||
| def arbitrary_speed_subsample(frames_speed, | ||
| num_steps: int, | ||
| random: bool, | ||
| img_size: int, | ||
| channels: int = 3, | ||
| stride: int = 1, | ||
| seed: Optional[int] = None) -> tf.Tensor: | ||
| frames, speed = frames_speed | ||
| frame_len = tf.cast(tf.shape(frames)[0], tf.float32) | ||
| max_frame_len = tf.math.floordiv(frame_len, speed) if speed > 1 else frame_len | ||
|
|
||
| x_ref = tf.range(0, speed * (len(frames) - 0.5), speed, dtype=tf.float32) | ||
| x_ref = tf.stack([x_ref] * (img_size * img_size * channels)) | ||
| new_frames = tfp.math.batch_interp_regular_1d_grid( | ||
| x=x_ref, | ||
| x_ref_min=[0] * (img_size * img_size * channels), | ||
| x_ref_max=[len(frames)] * (img_size * img_size * channels), | ||
| y_ref=tf.transpose(tf.reshape(frames, [len(frames), -1])) | ||
| ) | ||
| sequence = tf.reshape( | ||
| tf.transpose(new_frames), frames.shape.as_list() | ||
| )[:tf.cast(max_frame_len, tf.int32)] | ||
|
|
||
| sequence_length = tf.shape(sequence)[0] | ||
|
|
||
| if random: | ||
| sequence_length = tf.cast(sequence_length, tf.float32) | ||
| frame_stride = tf.cast(stride, tf.float32) | ||
| max_offset = tf.cond( | ||
| sequence_length > (num_steps - 1) * frame_stride, | ||
| lambda: sequence_length - (num_steps - 1) * frame_stride, | ||
| lambda: sequence_length) | ||
| offset = tf.random.uniform((), | ||
| maxval=tf.cast(max_offset, dtype=tf.int32), | ||
| dtype=tf.int32, | ||
| seed=seed) | ||
| else: | ||
| raise NotImplementedError(f"Only `random == True` is supported now.") | ||
| offset = (sequence_length - num_steps * stride) // 2 | ||
| offset = tf.maximum(0, offset) | ||
|
|
||
| indices = _sample_or_pad_sequence_indices( | ||
| sequence=sequence, num_steps=num_steps, stride=stride, offset=offset) | ||
| indices.set_shape((num_steps,)) | ||
|
|
||
| return tf.gather(sequence, indices) | ||
|
|
||
|
|
||
| # (batched) Arbitrary speed / frequency augmentation for SimPer | ||
| def batched_arbitrary_speed(frames, num_diff_speeds, speed_range=(0.5, 2)): | ||
| random_speeds = tf.random.uniform([num_diff_speeds], | ||
| minval=speed_range[0], | ||
| maxval=speed_range[1], | ||
| dtype=tf.float32) | ||
| random_speeds = tf.sort(random_speeds) | ||
| random_speeds = tf.concat([random_speeds, random_speeds], 0) | ||
|
|
||
| # construct (2 * M) sub-video batch for SimPer loss | ||
| batched_frames = tf.stack([frames] * num_diff_speeds * 2) | ||
| batched_frames = tf.map_fn( | ||
| arbitrary_speed_subsample, (batched_frames, random_speeds), | ||
| fn_output_signature=tf.float32) | ||
|
|
||
| return batched_frames, random_speeds |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 1,83 @@ | ||
| """ | ||
| Example network architectures: | ||
| - Featurizer (for representation learning) | ||
| - Classifier (for downstream tasks) | ||
| """ | ||
| import tensorflow as tf | ||
| from tensorflow.keras.layers import (Conv2D, Conv3D, Dense, Flatten, BatchNormalization, | ||
| TimeDistributed, MaxPool2D, GlobalAveragePooling2D) | ||
|
|
||
|
|
||
| class Featurizer(tf.keras.Model): | ||
|
|
||
| def __init__(self, n_outputs): | ||
| super(Featurizer, self).__init__() | ||
| self.conv0 = Conv3D(64, (5, 3, 3), padding='same') | ||
| self.conv1 = Conv3D(128, (5, 3, 3), padding='same') | ||
| self.conv2 = Conv3D(128, (5, 3, 3), padding='same') | ||
| self.conv3 = Conv3D(1, (1, 1, 1)) | ||
|
|
||
| self.bn0 = BatchNormalization() | ||
| self.bn1 = BatchNormalization() | ||
| self.bn2 = BatchNormalization() | ||
| self.bn3 = BatchNormalization() | ||
|
|
||
| self.pool0 = TimeDistributed(MaxPool2D((2, 2))) | ||
| self.pool1 = TimeDistributed(MaxPool2D((2, 2))) | ||
| self.pool2 = TimeDistributed(MaxPool2D((2, 2))) | ||
| self.pool3 = TimeDistributed(GlobalAveragePooling2D()) | ||
|
|
||
| self.flatten = Flatten() | ||
|
|
||
| self.n_outputs = n_outputs | ||
|
|
||
| def call(self, x): | ||
| x = self.conv0(x) | ||
| x = self.bn0(x) | ||
| x = tf.nn.relu(x) | ||
| x = self.pool0(x) | ||
|
|
||
| x = self.conv1(x) | ||
| x = self.bn1(x) | ||
| x = tf.nn.relu(x) | ||
| x = self.pool1(x) | ||
|
|
||
| # x = self.conv2(x) | ||
| # x = self.bn2(x) | ||
| # x = tf.nn.relu(x) | ||
| # x = self.pool2(x) | ||
|
|
||
| x = self.conv3(x) | ||
| x = self.bn3(x) | ||
| x = tf.nn.relu(x) | ||
| x = self.pool3(x) | ||
|
|
||
| x = self.flatten(x) | ||
| return x | ||
|
|
||
|
|
||
| class MLP(tf.keras.Model): | ||
|
|
||
| def __init__(self, n_outputs): | ||
| super(MLP, self).__init__() | ||
| self.inputs = Dense(n_outputs) | ||
| self.hidden = Dense(n_outputs) | ||
| self.outputs = Dense(n_outputs) | ||
|
|
||
| def call(self, x): | ||
| x = self.inputs(x) | ||
| x = tf.nn.relu(x) | ||
| # x = self.hidden(x) | ||
| # x = tf.nn.relu(x) | ||
| x = self.outputs(x) | ||
| return x | ||
|
|
||
|
|
||
| def Classifier(in_features, out_features, nonlinear=False): | ||
| if nonlinear: | ||
| return tf.keras.Sequential( | ||
| [Dense(in_features // 2, activation=tf.nn.relu), | ||
| Dense(in_features // 4, activation=tf.nn.relu), | ||
| Dense(out_features)]) | ||
| else: | ||
| return Dense(out_features) |
Oops, something went wrong.