Code
stringlengths 103
85.9k
| Summary
sequencelengths 0
94
|
|---|---|
Please provide a description of the function:def van_image_enc_2d(x, first_depth, reuse=False, hparams=None):
with tf.variable_scope('van_image_enc', reuse=reuse):
enc_history = [x]
enc = tf.layers.conv2d(
x, first_depth, 3, padding='same', activation=tf.nn.relu, strides=1)
enc = tf.contrib.layers.layer_norm(enc)
enc = tf.layers.conv2d(
enc, first_depth, 3, padding='same', activation=tf.nn.relu, strides=1)
enc = tf.nn.max_pool(enc, [1, 2, 2, 1], [1, 2, 2, 1], 'SAME')
enc = tf.nn.dropout(enc, hparams.van_keep_prob)
enc = tf.contrib.layers.layer_norm(enc)
enc_history.append(enc)
enc = tf.layers.conv2d(
enc,
first_depth * 2,
3,
padding='same',
activation=tf.nn.relu,
strides=1)
enc = tf.layers.conv2d(
enc,
first_depth * 2,
3,
padding='same',
activation=tf.nn.relu,
strides=1)
enc = tf.nn.max_pool(enc, [1, 2, 2, 1], [1, 2, 2, 1], 'SAME')
enc = tf.nn.dropout(enc, hparams.van_keep_prob)
enc = tf.contrib.layers.layer_norm(enc)
enc_history.append(enc)
enc = tf.layers.conv2d(
enc,
first_depth * 4,
3,
padding='same',
activation=tf.nn.relu,
strides=1)
enc = tf.layers.conv2d(
enc,
first_depth * 4,
3,
padding='same',
activation=tf.nn.relu,
strides=1)
enc = tf.layers.conv2d(
enc,
first_depth * 4,
3,
padding='same',
activation=tf.nn.relu,
strides=1)
enc = tf.nn.max_pool(enc, [1, 2, 2, 1], [1, 2, 2, 1], 'SAME')
return enc, enc_history | [
"The image encoder for the VAN.\n\n Similar architecture as Ruben's paper\n (http://proceedings.mlr.press/v70/villegas17a/villegas17a.pdf).\n\n Args:\n x: The image to encode.\n first_depth: The depth of the first layer. Depth is increased in subsequent\n layers.\n reuse: To reuse in variable scope or not.\n hparams: The python hparams.\n\n Returns:\n The encoded image.\n "
] |
Please provide a description of the function:def van_enc_2d(x, first_depth, reuse=False):
with tf.variable_scope('van_enc', reuse=reuse):
a = 4 # depends on the inputs size
b = 4
# a, b = 4,4
enc = tf.nn.relu(x)
enc = tf.layers.dense(enc, first_depth * a * b, tf.nn.relu)
enc = tf.contrib.layers.layer_norm(enc)
enc = tf.reshape(enc, [-1, a, b, first_depth])
enc = tf.layers.conv2d_transpose(
enc, first_depth, 3, padding='same', activation=tf.nn.relu, strides=1)
enc = tf.contrib.layers.layer_norm(enc)
enc = tf.layers.conv2d_transpose(
enc,
first_depth * 2,
3,
padding='same',
activation=tf.nn.relu,
strides=2)
van_higher_level_2 = tf.reshape(enc, [-1, a * 2 * b * 2 * first_depth * 2])
enc = tf.layers.conv2d_transpose(
enc,
first_depth * 2,
3,
padding='same',
activation=tf.nn.relu,
strides=1)
enc = tf.contrib.layers.layer_norm(enc)
enc = tf.layers.conv2d_transpose(
enc,
first_depth * 4,
3,
padding='same',
activation=tf.nn.relu,
strides=1)
van_higher_level_4 = tf.reshape(enc, [-1, a * 2 * b * 2 * first_depth * 4])
van_higher_level = tf.concat([x, van_higher_level_2, van_higher_level_4], 1)
return enc, van_higher_level | [
"The higher level structure encoder for the VAN.\n\n The high level structure is a vector instead of an image.\n\n Args:\n x: The higher level structure to encode.\n first_depth: The depth of the first layer. Depth is increased in subsequent\n layers.\n reuse: To reuse in variable scope or not.\n\n Returns:\n The encoded image.\n "
] |
Please provide a description of the function:def van_dec_2d(x, skip_connections, output_shape, first_depth, hparams=None):
with tf.variable_scope('van_dec'):
dec = tf.layers.conv2d_transpose(
x, first_depth * 4, 3, padding='same', activation=tf.nn.relu, strides=2)
dec = tf.nn.dropout(dec, hparams.van_keep_prob)
dec = tf.contrib.layers.layer_norm(dec)
dec = tf.layers.conv2d_transpose(
dec,
first_depth * 4,
3,
padding='same',
activation=tf.nn.relu,
strides=1)
dec = tf.nn.dropout(dec, hparams.van_keep_prob)
dec = tf.layers.conv2d_transpose(
dec,
first_depth * 2,
3,
padding='same',
activation=tf.nn.relu,
strides=1)
dec = tf.nn.dropout(dec, hparams.van_keep_prob)
dec = tf.contrib.layers.layer_norm(dec)
dec = tf.layers.conv2d_transpose(
dec,
first_depth * 2,
3,
padding='same',
activation=tf.nn.relu,
strides=2)
dec = tf.nn.dropout(dec, hparams.van_keep_prob)
dec = tf.layers.conv2d_transpose(
dec, first_depth, 3, padding='same', activation=tf.nn.relu, strides=1)
dec = tf.nn.dropout(dec, hparams.van_keep_prob)
dec = tf.contrib.layers.layer_norm(dec)
dec = tf.layers.conv2d_transpose(
dec,
output_shape[3] + 1,
3,
padding='same',
activation=tf.nn.relu,
strides=2)
dec = tf.nn.dropout(dec, hparams.van_keep_prob)
out_mask = tf.layers.conv2d_transpose(
dec, output_shape[3] + 1, 3, strides=1, padding='same', activation=None)
mask = tf.nn.sigmoid(out_mask[:, :, :, 3:4])
out = out_mask[:, :, :, :3]
return out * mask + skip_connections[0] * (1 - mask) | [
"The VAN decoder.\n\n Args:\n x: The analogy information to decode.\n skip_connections: The encoder layers which can be used as skip connections.\n output_shape: The shape of the desired output image.\n first_depth: The depth of the first layer of the van image encoder.\n hparams: The python hparams.\n\n Returns:\n The decoded image prediction.\n "
] |
Please provide a description of the function:def analogy_computation_2d(f_first_enc,
f_first_frame,
f_current_enc,
first_depth):
with tf.variable_scope('analogy_computation'):
frame_enc_diff = f_first_frame - f_first_enc
frame_enc_diff_enc = tf.layers.conv2d(
frame_enc_diff,
first_depth * 4,
3,
padding='same',
activation=tf.nn.relu,
strides=1)
f_current_enc_enc = tf.layers.conv2d(
f_current_enc,
first_depth * 4,
3,
padding='same',
activation=tf.nn.relu,
strides=1)
analogy = tf.concat([frame_enc_diff_enc, f_current_enc_enc], 3)
analogy = tf.layers.conv2d(
analogy,
first_depth * 4,
3,
padding='same',
activation=tf.nn.relu,
strides=1)
analogy = tf.contrib.layers.layer_norm(analogy)
analogy = tf.layers.conv2d(
analogy,
first_depth * 4,
3,
padding='same',
activation=tf.nn.relu,
strides=1)
return tf.layers.conv2d(
analogy,
first_depth * 4,
3,
padding='same',
activation=tf.nn.relu,
strides=1) | [
"Implements the deep analogy computation."
] |
Please provide a description of the function:def van(first_enc,
first_frame,
current_enc,
gt_image,
reuse=False,
scope_prefix='',
hparams=None):
with tf.variable_scope(scope_prefix + 'van', reuse=reuse):
output_shape = first_frame.get_shape().as_list()
output_shape[0] = -1
first_depth = 64
f_first_enc, _ = van_enc_2d(first_enc, first_depth)
f_first_frame, image_enc_history = van_image_enc_2d(
first_frame, first_depth, hparams=hparams)
f_current_enc, van_higher_level = van_enc_2d(
current_enc, first_depth, reuse=True)
f_gt_image, _ = van_image_enc_2d(gt_image, first_depth, True,
hparams=hparams)
analogy_t = analogy_computation_2d(
f_first_enc, f_first_frame, f_current_enc, first_depth)
enc_img = f_current_enc + analogy_t
img = van_dec_2d(
enc_img, image_enc_history, output_shape, first_depth, hparams=hparams)
batch_size = tf.to_float(tf.shape(first_enc)[0])
r_loss = tf.nn.l2_loss(f_gt_image - f_current_enc - analogy_t) / batch_size
return img, r_loss, van_higher_level | [
"Implements a VAN.\n\n Args:\n first_enc: The first encoding.\n first_frame: The first ground truth frame.\n current_enc: The encoding of the frame to generate.\n gt_image: The ground truth image, only used for regularization.\n reuse: To reuse in variable scope or not.\n scope_prefix: The prefix before the scope name.\n hparams: The python hparams.\n\n Returns:\n The generated image.\n "
] |
Please provide a description of the function:def encoder_vgg(x, enc_final_size, reuse=False, scope_prefix='', hparams=None,
is_training=True):
with tf.variable_scope(scope_prefix + 'encoder', reuse=reuse):
# Preprocess input
x *= 256
x = x - COLOR_NORMALIZATION_VECTOR
with arg_scope(vgg.vgg_arg_scope()):
# Padding because vgg_16 accepts images of size at least VGG_IMAGE_SIZE.
x = tf.pad(x, [[0, 0], [0, VGG_IMAGE_SIZE - IMG_WIDTH],
[0, VGG_IMAGE_SIZE - IMG_HEIGHT], [0, 0]])
_, end_points = vgg.vgg_16(
x,
num_classes=enc_final_size,
is_training=is_training)
pool5_key = [key for key in end_points.keys() if 'pool5' in key]
assert len(pool5_key) == 1
enc = end_points[pool5_key[0]]
# Undoing padding.
enc = tf.slice(enc, [0, 0, 0, 0], [-1, 2, 2, -1])
enc_shape = enc.get_shape().as_list()
enc_shape[0] = -1
enc_size = enc_shape[1] * enc_shape[2] * enc_shape[3]
enc_flat = tf.reshape(enc, (-1, enc_size))
enc_flat = tf.nn.dropout(enc_flat, hparams.enc_keep_prob)
enc_flat = tf.layers.dense(
enc_flat,
enc_final_size,
kernel_initializer=tf.truncated_normal_initializer(stddev=1e-4,))
if hparams.enc_pred_use_l2norm:
enc_flat = tf.nn.l2_normalize(enc_flat, 1)
return enc_flat | [
"VGG network to use as encoder without the top few layers.\n\n Can be pretrained.\n\n Args:\n x: The image to encode. In the range 0 to 1.\n enc_final_size: The desired size of the encoding.\n reuse: To reuse in variable scope or not.\n scope_prefix: The prefix before the scope name.\n hparams: The python hparams.\n is_training: boolean value indicating if training is happening.\n\n Returns:\n The generated image.\n "
] |
Please provide a description of the function:def predictor(enc_flat,
action,
lstm_states,
pred_depth,
reuse=False,
scope_prefix='',
hparams=None):
with tf.variable_scope(scope_prefix + 'predict', reuse=reuse):
enc_final_size = enc_flat.get_shape().as_list()[1]
action_size = action.get_shape().as_list()[1]
initial_size = (enc_final_size + action_size)
batch_size = tf.shape(enc_flat)[0]
init_stddev = 1e-2
pre_pred = tf.concat([enc_flat, action], 1)
pre_pred = tf.layers.dense(
pre_pred,
initial_size,
kernel_initializer=tf.truncated_normal_initializer(stddev=init_stddev))
# This is only needed or the GAN version.
if hparams.pred_noise_std > 0:
# Add the noise like this so a pretrained model can be used.
pred_noise = tf.random_normal(
shape=[batch_size, 100], stddev=hparams.pred_noise_std)
pre_pred += tf.layers.dense(
pred_noise,
initial_size,
kernel_initializer=tf.truncated_normal_initializer(
stddev=init_stddev),
name='noise_dense')
pre_pred = tf.nn.relu(pre_pred)
if lstm_states[pred_depth - 2] is None:
back_connect = tf.tile(
tf.get_variable(
'back_connect_init',
shape=[1, initial_size * 2],
initializer=tf.truncated_normal_initializer(stddev=init_stddev))
, (batch_size, 1))
else:
back_connect = lstm_states[pred_depth - 2]
lstm_init_stddev = 1e-4
part_pred, lstm_states[0] = common_video.lstm_cell(
tf.concat([pre_pred, back_connect], 1),
lstm_states[0],
initial_size,
use_peepholes=True,
initializer=tf.truncated_normal_initializer(stddev=lstm_init_stddev),
num_proj=initial_size)
part_pred = tf.contrib.layers.layer_norm(part_pred)
pred = part_pred
for pred_layer_num in range(1, pred_depth, 2):
part_pred, lstm_states[pred_layer_num] = common_video.lstm_cell(
pred,
lstm_states[pred_layer_num],
initial_size,
use_peepholes=True,
initializer=tf.truncated_normal_initializer(stddev=lstm_init_stddev),
num_proj=initial_size)
pred += part_pred
part_pred, lstm_states[pred_layer_num + 1] = common_video.lstm_cell(
tf.concat([pred, pre_pred], 1),
lstm_states[pred_layer_num + 1],
initial_size,
use_peepholes=True,
initializer=tf.truncated_normal_initializer(stddev=lstm_init_stddev),
num_proj=initial_size)
part_pred = tf.contrib.layers.layer_norm(part_pred)
pred += part_pred
pred = tf.layers.dense(
pred,
enc_final_size,
kernel_initializer=tf.truncated_normal_initializer(stddev=init_stddev))
if hparams.enc_pred_use_l2norm:
pred = tf.nn.l2_normalize(pred, 1)
return pred | [
"LSTM predictor network."
] |
Please provide a description of the function:def construct_model(images,
actions=None,
context_frames=2,
hparams=None,
is_training=True):
pred_depth = 20
enc_out_all, pred_out_all, van_out_all, van_on_enc_all = [], [], [], []
lstm_states = [None] * (pred_depth + 2)
enc_out = encoder_vgg(
images[0], hparams.enc_size, False, scope_prefix='timestep/',
hparams=hparams, is_training=is_training)
enc_out = tf.identity(enc_out, 'enc_out')
enc_out_all.append(enc_out)
num_timesteps = len(actions) - 1
sum_freq = int(num_timesteps / 4 + 1)
reuse = False
for timestep, action in zip(range(len(actions) - 1), actions[:-1]):
done_warm_start = timestep > context_frames - 1
with tf.variable_scope('timestep', reuse=reuse):
if done_warm_start:
pred_input = pred_out_all[-1]
else:
pred_input = enc_out_all[-1]
pred_out = predictor(
pred_input, action, lstm_states, pred_depth, False, hparams=hparams)
pred_out = tf.identity(pred_out, 'pred_out')
if timestep % sum_freq == 0: # and not hparams.use_tpu:
tf.summary.histogram('pred_out', pred_out)
pred_out_all.append(pred_out)
if timestep % sum_freq == 0: # and not hparams.use_tpu:
tf.summary.histogram('lstm_state', lstm_states[0])
van_out, _, _ = van(
enc_out_all[0],
images[0],
pred_out,
images[timestep + 1],
tf.AUTO_REUSE,
hparams=hparams)
van_out = tf.identity(van_out, 'van_out')
van_out_all.append(van_out)
enc_out = encoder_vgg(
images[timestep + 1], hparams.enc_size, True, hparams=hparams,
is_training=is_training)
enc_out = tf.identity(enc_out, 'enc_out')
if timestep % sum_freq == 0: # and not hparams.use_tpu:
tf.summary.histogram('enc_out', enc_out)
enc_out_all.append(enc_out)
van_input = images[0]
enc_noise = tf.zeros_like(enc_out)
if timestep % sum_freq == 0: # and not hparams.use_tpu:
tf.summary.histogram('enc_noise', enc_noise)
van_on_enc, _, _ = van(
enc_out_all[0],
van_input,
enc_out + enc_noise,
images[timestep + 1],
tf.AUTO_REUSE,
hparams=hparams)
van_on_enc = tf.identity(van_on_enc, 'van_on_enc')
van_on_enc_all.append(van_on_enc)
reuse = True
return enc_out_all, pred_out_all, van_out_all, van_on_enc_all | [
"Constructs the tensorflow graph of the hierarchical model."
] |
Please provide a description of the function:def peak_signal_to_noise_ratio(true, pred):
return 10.0 * tf.log(1.0 / mean_squared_error(true, pred)) / tf.log(10.0) | [
"Image quality metric based on maximal signal power vs. power of the noise.\n\n Args:\n true: the ground truth image.\n pred: the predicted image.\n Returns:\n peak signal to noise ratio (PSNR)\n "
] |
Please provide a description of the function:def mean_squared_error(true, pred):
result = tf.reduce_sum(
tf.squared_difference(true, pred)) / tf.to_float(tf.size(pred))
return result | [
"L2 distance between tensors true and pred.\n\n Args:\n true: the ground truth image.\n pred: the predicted image.\n Returns:\n mean squared error between ground truth and predicted image.\n "
] |
Please provide a description of the function:def l1_error(true, pred):
return tf.reduce_sum(tf.abs(true - pred)) / tf.to_float(tf.size(pred)) | [
"L1 distance between tensors true and pred."
] |
Please provide a description of the function:def calc_loss_psnr(gen_images, images, name, hparams=None, use_l1_loss=False):
del hparams
with tf.name_scope(name):
loss, error, psnr_all = 0.0, 0.0, 0.0
for _, x, gx in zip(range(len(gen_images)), images, gen_images):
recon_cost = mean_squared_error(x, gx)
if use_l1_loss:
recon_cost = l1_error(x, gx)
error_i = l1_error(x, gx)
psnr_i = peak_signal_to_noise_ratio(x, gx)
psnr_all += psnr_i
error += error_i
loss += recon_cost
psnr_all /= tf.to_float(len(gen_images))
loss /= tf.to_float(len(gen_images))
error /= tf.to_float(len(gen_images))
# if not hparams.use_tpu:
tf.summary.scalar('psnr_all', psnr_all)
tf.summary.scalar('loss', loss)
return loss, psnr_all | [
"Calculates loss and psnr for predictions over multiple timesteps."
] |
Please provide a description of the function:def next_frame_sv2p():
hparams = basic_stochastic.next_frame_basic_stochastic()
hparams.optimizer = "true_adam"
hparams.learning_rate_schedule = "constant"
hparams.learning_rate_constant = 1e-3
hparams.video_num_input_frames = 1
hparams.video_num_target_frames = 3
hparams.batch_size = 16
hparams.bottom = {
"inputs": modalities.video_raw_bottom,
"targets": modalities.video_raw_targets_bottom,
}
hparams.loss = {
"targets": modalities.video_l2_raw_loss,
}
hparams.top = {
"targets": modalities.video_raw_top,
}
hparams.video_modality_loss_cutoff = 0.0
hparams.scheduled_sampling_mode = "count"
hparams.scheduled_sampling_k = 900.0
hparams.add_hparam("reward_prediction", True)
hparams.add_hparam("reward_prediction_stop_gradient", False)
hparams.add_hparam("reward_prediction_buffer_size", 0)
hparams.add_hparam("model_options", "CDNA")
hparams.add_hparam("num_masks", 10)
hparams.add_hparam("multi_latent", False)
hparams.add_hparam("relu_shift", 1e-12)
hparams.add_hparam("dna_kernel_size", 5)
hparams.add_hparam("upsample_method", "conv2d_transpose")
hparams.add_hparam("reward_model", "basic")
hparams.add_hparam("visualize_logits_histogram", True)
return hparams | [
"SV2P model hparams."
] |
Please provide a description of the function:def next_frame_sv2p_discrete():
hparams = next_frame_sv2p()
hparams.action_injection = "multiplicative"
hparams.small_mode = True
hparams.add_hparam("bottleneck_bits", 128)
hparams.add_hparam("bottleneck_noise", 0.02)
hparams.add_hparam("discrete_warmup_steps", 40000)
hparams.add_hparam("full_latent_tower", False)
hparams.add_hparam("latent_predictor_state_size", 128)
hparams.add_hparam("latent_predictor_temperature", 0.5)
hparams.add_hparam("discretize_warmup_steps", 40000)
return hparams | [
"SV2P discrete model hparams."
] |
Please provide a description of the function:def next_frame_sv2p_atari():
hparams = next_frame_sv2p()
hparams.video_num_input_frames = 4
hparams.video_num_target_frames = 4
hparams.action_injection = "multiplicative"
hparams.num_iterations_1st_stage = 12000
hparams.num_iterations_2nd_stage = 12000
hparams.anneal_end = 40000
hparams.latent_loss_multiplier_schedule = "noisy_linear_cosine_decay"
hparams.latent_loss_multiplier = 1e-3
hparams.information_capacity = 0.0
hparams.small_mode = True
return hparams | [
"SV2P model for atari."
] |
Please provide a description of the function:def next_frame_sv2p_atari_softmax():
hparams = next_frame_sv2p_atari()
hparams.bottom = {}
hparams.loss = {}
hparams.top = {}
hparams.internal_loss = True
return hparams | [
"SV2P model for atari with softmax."
] |
Please provide a description of the function:def next_frame_sv2p_tiny():
hparams = next_frame_sv2p_atari_softmax()
hparams.batch_size = 2
hparams.tiny_mode = True
hparams.num_masks = 1
hparams.video_modality_loss_cutoff = 0.4
hparams.video_num_input_frames = 4
hparams.video_num_target_frames = 4
return hparams | [
"Tiny SV2P model."
] |
Please provide a description of the function:def next_frame_sv2p_cutoff():
hparams = next_frame_sv2p()
hparams.video_modality_loss_cutoff = 0.4
hparams.video_num_input_frames = 4
hparams.video_num_target_frames = 1
return hparams | [
"SV2P model with additional cutoff in L2 loss for environments like pong."
] |
Please provide a description of the function:def _get_mscoco(directory):
for url in _MSCOCO_URLS:
filename = os.path.basename(url)
download_url = os.path.join(_MSCOCO_ROOT_URL, url)
path = generator_utils.maybe_download(directory, filename, download_url)
unzip_dir = os.path.join(directory, filename.strip(".zip"))
if not tf.gfile.Exists(unzip_dir):
zipfile.ZipFile(path, "r").extractall(directory) | [
"Download and extract MSCOCO datasets to directory unless it is there."
] |
Please provide a description of the function:def mscoco_generator(data_dir,
tmp_dir,
training,
how_many,
start_from=0,
eos_list=None,
vocab_filename=None):
eos_list = [1] if eos_list is None else eos_list
def get_vocab():
if data_dir is not None and vocab_filename is not None:
vocab_filepath = os.path.join(data_dir, vocab_filename)
if tf.gfile.Exists(vocab_filepath):
tf.logging.info("Found vocab file: %s", vocab_filepath)
vocab_symbolizer = text_encoder.SubwordTextEncoder(vocab_filepath)
return vocab_symbolizer
else:
raise ValueError("Vocab file does not exist: %s" % vocab_filepath)
return None
vocab_symbolizer = get_vocab()
_get_mscoco(tmp_dir)
caption_filepath = (
_MSCOCO_TRAIN_CAPTION_FILE if training else _MSCOCO_EVAL_CAPTION_FILE)
caption_filepath = os.path.join(tmp_dir, caption_filepath)
prefix = _MSCOCO_TRAIN_PREFIX if training else _MSCOCO_EVAL_PREFIX
caption_file = io.open(caption_filepath)
caption_json = json.load(caption_file)
# Dictionary from image_id to ((filename, height, width), captions).
image_dict = {}
for image in caption_json["images"]:
image_dict[image["id"]] = [(image["file_name"], image["height"],
image["width"]), []]
annotations = caption_json["annotations"]
annotation_count = len(annotations)
image_count = len(image_dict)
tf.logging.info("Processing %d images and %d labels\n" % (image_count,
annotation_count))
for annotation in annotations:
image_id = annotation["image_id"]
image_dict[image_id][1].append(annotation["caption"])
data = list(image_dict.values())[start_from:start_from + how_many]
random.shuffle(data)
for image_info, labels in data:
image_filename = image_info[0]
image_filepath = os.path.join(tmp_dir, prefix, image_filename)
with tf.gfile.Open(image_filepath, "rb") as f:
encoded_image_data = f.read()
height, width = image_info[1], image_info[2]
for label in labels:
if vocab_filename is None or vocab_symbolizer is None:
label = [ord(c) for c in label] + eos_list
else:
label = vocab_symbolizer.encode(label) + eos_list
yield {
"image/encoded": [encoded_image_data],
"image/format": ["jpeg"],
"image/class/label": label,
"image/height": [height],
"image/width": [width]
} | [
"Image generator for MSCOCO captioning problem with token-wise captions.\n\n Args:\n data_dir: path to the data directory.\n tmp_dir: path to temporary storage directory.\n training: a Boolean; if true, we use the train set, otherwise the test set.\n how_many: how many images and labels to generate.\n start_from: from which image to start.\n eos_list: optional list of end of sentence tokens, otherwise use default\n value `1`.\n vocab_filename: file within `tmp_dir` to read vocabulary from.\n\n Yields:\n A dictionary representing the images with the following fields:\n * image/encoded: the string encoding the image as JPEG,\n * image/format: the string \"jpeg\" representing image format,\n * image/class/label: a list of integers representing the caption,\n * image/height: an integer representing the height,\n * image/width: an integer representing the width.\n Every field is actually a list of the corresponding type.\n ",
"Get vocab for caption text encoder."
] |
Please provide a description of the function:def flags_as_args():
if hasattr(FLAGS, "flag_values_dict"):
args_dict = FLAGS.flag_values_dict()
else:
args_dict = dict(FLAGS.__dict__["__flags"])
del args_dict["cloud_mlengine"]
# Configured later
del args_dict["t2t_usr_dir"]
args_dict.pop("h", None)
args_dict.pop("helpfull", None)
args_dict.pop("helpshort", None)
args_dict.pop("help", None)
args = []
for name, val in args_dict.items():
if val is None:
continue
if name.startswith("autotune"):
continue
args.extend(["--%s=%s" % (name, str(val))])
return args | [
"Convert FLAGS to list of args suitable for passing on cmd line."
] |
Please provide a description of the function:def get_default_master_type(num_gpus=1):
gpus_to_master_map = {
0: "standard",
1: "standard_p100",
4: "complex_model_m_p100",
8: "complex_model_l_gpu",
}
if num_gpus not in gpus_to_master_map:
raise ValueError("Num gpus must be in %s" %
str(sorted(list(gpus_to_master_map.keys()))))
return gpus_to_master_map[num_gpus] | [
"Returns master_type for trainingInput."
] |
Please provide a description of the function:def configure_job():
# See documentation:
# https://cloud.google.com/ml-engine/reference/rest/v1/projects.jobs#traininginput
training_input = {
"pythonModule": "tensor2tensor.bin.t2t_trainer",
"args": flags_as_args(),
"region": text_encoder.native_to_unicode(default_region()),
"runtimeVersion": RUNTIME_VERSION,
"pythonVersion": "3.5" if sys.version_info.major == 3 else "2.7",
"jobDir": FLAGS.output_dir,
"scaleTier": "CUSTOM",
"masterType": FLAGS.cloud_mlengine_master_type or get_default_master_type(
num_gpus=FLAGS.worker_gpu)
}
if FLAGS.use_tpu:
training_input["masterType"] = (FLAGS.cloud_mlengine_master_type or
"standard")
training_input["workerType"] = "cloud_tpu"
training_input["workerCount"] = 1
if FLAGS.hparams_range:
tf.logging.info("Configuring hyperparameter tuning.")
training_input["hyperparameters"] = configure_autotune(
FLAGS.hparams_range,
FLAGS.autotune_objective,
FLAGS.autotune_maximize,
FLAGS.autotune_max_trials,
FLAGS.autotune_parallel_trials,
)
timestamp = datetime.datetime.now().strftime("%Y%m%d_%H%M%S")
job_spec = {
"jobId": "%s_%s_t2t_%s" % (FLAGS.model, FLAGS.problem, timestamp),
"labels": {
"model": FLAGS.model,
"problem": FLAGS.problem,
"hparams": FLAGS.hparams_set
},
"trainingInput": training_input,
}
return job_spec | [
"Construct jobSpec for ML Engine job."
] |
Please provide a description of the function:def launch_job(job_spec):
project_id = "projects/{}".format(
text_encoder.native_to_unicode(default_project()))
credentials = GoogleCredentials.get_application_default()
cloudml = discovery.build("ml", "v1", credentials=credentials,
cache_discovery=False)
request = cloudml.projects().jobs().create(body=job_spec, parent=project_id)
request.execute() | [
"Launch job on ML Engine."
] |
Please provide a description of the function:def _tar_and_copy(src_dir, target_dir):
src_dir = src_dir.rstrip("/")
target_dir = target_dir.rstrip("/")
tmp_dir = tempfile.gettempdir().rstrip("/")
src_base = os.path.basename(src_dir)
shell_run(
"tar --exclude=.git -zcf {tmp_dir}/{src_base}.tar.gz -C {src_dir} .",
src_dir=src_dir,
src_base=src_base,
tmp_dir=tmp_dir)
final_destination = "%s/%s.tar.gz" % (target_dir, src_base)
shell_run(
("gsutil cp {tmp_dir}/{src_base}.tar.gz "
"{final_destination}"),
tmp_dir=tmp_dir,
src_base=src_base,
final_destination=final_destination)
return final_destination | [
"Tar and gzip src_dir and copy to GCS target_dir."
] |
Please provide a description of the function:def tar_and_copy_t2t(train_dir):
tf.logging.info("Tarring and pushing local Tensor2Tensor package.")
output = text_encoder.native_to_unicode(shell_output(
"pip show tensor2tensor")).split("\n")
assert output[1].startswith("Version")
assert output[7].startswith("Location")
t2t_version = output[1].split(":")[1].strip()
t2t_dir = output[7].split(":")[1].strip()
# A local installation cloned from GitHub will have a setup.py file and a docs
# folder
is_local_t2t = all([
tf.gfile.Exists(os.path.join(t2t_dir, fname))
for fname in ["setup.py", "docs/cloud_mlengine.md"]
])
if is_local_t2t:
tf.logging.info("Found local T2T installation. Tarring directory %s",
t2t_dir)
else:
# PyPI installation
# Create a folder with just a setup.py file pointing to the right version
tf.logging.info("Found PyPI T2T installation. Launching tensor2tensor==%s",
t2t_version)
t2t_dir = os.path.join(tempfile.gettempdir(), "tensor2tensor_tmp")
shutil.rmtree(t2t_dir, ignore_errors=True)
os.mkdir(t2t_dir)
setup_fname = os.path.join(t2t_dir, "setup.py")
setup_file_str = get_setup_file(
name="DummyT2TPackage",
packages=["tensor2tensor==%s" % t2t_version]
)
with tf.gfile.Open(setup_fname, "w") as f:
f.write(setup_file_str)
t2t_tar = _tar_and_copy(t2t_dir, train_dir)
return t2t_tar | [
"Tar Tensor2Tensor and cp to train_dir."
] |
Please provide a description of the function:def tar_and_copy_usr_dir(usr_dir, train_dir):
tf.logging.info("Tarring and pushing t2t_usr_dir.")
usr_dir = os.path.abspath(os.path.expanduser(usr_dir))
# Copy usr dir to a temp location
top_dir = os.path.join(tempfile.gettempdir(), "t2t_usr_container")
tmp_usr_dir = os.path.join(top_dir, usr_dir_lib.INTERNAL_USR_DIR_PACKAGE)
shutil.rmtree(top_dir, ignore_errors=True)
shutil.copytree(usr_dir, tmp_usr_dir)
# Insert setup.py if one does not exist
top_setup_fname = os.path.join(top_dir, "setup.py")
setup_file_str = get_setup_file(
name="DummyUsrDirPackage",
packages=get_requirements(usr_dir)
)
with tf.gfile.Open(top_setup_fname, "w") as f:
f.write(setup_file_str)
usr_tar = _tar_and_copy(top_dir, train_dir)
return usr_tar | [
"Package, tar, and copy usr_dir to GCS train_dir."
] |
Please provide a description of the function:def validate_flags():
assert not job_dir()
assert FLAGS.output_dir.startswith("gs://")
assert FLAGS.data_dir.startswith("gs://")
assert FLAGS.worker_replicas <= 1
assert FLAGS.ps_replicas <= 0
if FLAGS.hparams_range:
assert FLAGS.autotune_objective
if FLAGS.worker_gpu:
assert FLAGS.worker_gpu in [1, 4, 8]
if FLAGS.cloud_mlengine_master_type:
if FLAGS.worker_gpu:
if FLAGS.worker_gpu == 1:
assert FLAGS.cloud_mlengine_master_type in ["standard_gpu",
"standard_p100"]
elif FLAGS.worker_gpu == 4:
assert FLAGS.cloud_mlengine_master_type in ["complex_model_m_gpu",
"complex_model_m_p100"]
else:
assert FLAGS.cloud_mlengine_master_type == "complex_model_l_gpu"
else:
assert FLAGS.cloud_mlengine_master_type in ["standard", "large_model",
"complex_model_s",
"complex_model_m",
"complex_model_l"] | [
"Validates flags are set to acceptable values for CloudML Engine runs."
] |
Please provide a description of the function:def launch():
validate_flags()
job_spec = configure_job()
job_name = job_spec["jobId"]
tf.logging.info("Launching job %s with ML Engine spec:\n%s", job_name,
pprint.pformat(job_spec))
assert confirm()
train_dir = FLAGS.output_dir
t2t_tar = tar_and_copy_t2t(train_dir)
configure_trainer_package(job_spec, t2t_tar)
if FLAGS.t2t_usr_dir:
usr_tar = tar_and_copy_usr_dir(FLAGS.t2t_usr_dir, train_dir)
configure_usr_dir(job_spec, usr_tar)
launch_job(job_spec)
tf.logging.info("Launched %s. See console to track: %s.", job_name,
CONSOLE_URL)
tf.logging.info("Interact with the training job from the command line:")
tf.logging.info("Abort job: gcloud ml-engine jobs cancel %s", job_name)
tf.logging.info("Stream logs: gcloud ml-engine jobs stream-logs %s", job_name)
tf.logging.info("Open tensorboard: tensorboard --logdir %s", train_dir) | [
"Launch t2t_trainer on Cloud ML Engine."
] |
Please provide a description of the function:def add_weight(cls):
@functools.wraps(cls.add_weight)
def _add_weight(self,
name=None,
shape=None,
dtype=None,
initializer=None,
regularizer=None,
**kwargs):
if isinstance(initializer, tf.keras.layers.Layer):
weight = initializer(shape, dtype)
self._trainable_weights.extend(initializer.trainable_weights) # pylint: disable=protected-access
self._non_trainable_weights.extend(initializer.non_trainable_weights) # pylint: disable=protected-access
if regularizer is not None:
# TODO(trandustin): Replace need for this with
# Layer._handle_weight_regularization. For Eager compatibility, random
# variable __init__s cannot apply TF ops (cl/220898007).
def loss_fn():
with tf.name_scope(name + '/Regularizer'):
return regularizer(initializer(shape, dtype))
self.add_loss(loss_fn)
return weight
return super(cls, self).add_weight(name=name,
shape=shape,
dtype=dtype,
initializer=initializer,
regularizer=regularizer,
**kwargs)
cls.add_weight = _add_weight
return cls | [
"Decorator for Layers, overriding add_weight for trainable initializers.",
"Adds weight.",
"Creates a regularization loss `Tensor`."
] |
Please provide a description of the function:def get_beta(self, kl_loss=0.0):
if self.hparams.latent_loss_multiplier_dynamic:
beta = tf.Variable(self.hparams.latent_loss_multiplier,
trainable=False, dtype=tf.float32)
alpha = self.hparams.latent_loss_multiplier_alpha
epsilon = self.hparams.latent_loss_multiplier_epsilon
shadow_beta = beta + alpha * (kl_loss - epsilon)
# Caping the beta between 0 and 1. May need to change this later on.
shadow_beta = tf.maximum(shadow_beta, 0.0)
shadow_beta = tf.minimum(shadow_beta, 1.0)
update_op = tf.assign(beta, shadow_beta)
else:
beta = common_video.beta_schedule(
schedule=self.hparams.latent_loss_multiplier_schedule,
global_step=self.get_iteration_num(),
final_beta=self.hparams.latent_loss_multiplier,
decay_start=(self.hparams.num_iterations_1st_stage +
self.hparams.num_iterations_2nd_stage),
decay_end=self.hparams.anneal_end)
update_op = tf.identity(beta) # fake update for regular beta.
with tf.control_dependencies([update_op]):
tf.summary.scalar("beta", beta)
return beta | [
"Get the KL multiplier, either dynamically or schedule based.\n\n if hparams.latent_loss_multiplier_dynamic is set to true, then beta\n is being adjusted to keep KL under hparams.latent_loss_multiplier_epsilon.\n In order to do so, the beta is being updated at each iteration\n by taking steps of size hparams.latent_loss_multiplier_alpha.\n The same formulation can be retrieved by solving the Lagrangian\n with KL < epsilon as a constraint.\n\n Args:\n kl_loss: KL loss. Only used for dynamic adjustment.\n\n Returns:\n beta: the final value of beta.\n\n "
] |
Please provide a description of the function:def get_kl_loss(self, means, log_vars, means_p=None, log_vars_p=None):
kl_loss = 0.0
if means_p is None:
means_p = tf.unstack(tf.zeros_like(means))
if log_vars_p is None:
log_vars_p = tf.unstack(tf.zeros_like(log_vars))
enumerated_inputs = enumerate(zip(means, log_vars, means_p, log_vars_p))
if self.is_training and self.hparams.stochastic_model:
for i, (mean, log_var, mean_p, log_var_p) in enumerated_inputs:
kl_loss += common_layers.kl_divergence(mean, log_var, mean_p, log_var_p)
tf.summary.histogram("posterior_mean_%d" % i, mean)
tf.summary.histogram("posterior_log_var_%d" % i, log_var)
tf.summary.histogram("prior_mean_%d" % i, mean_p)
tf.summary.histogram("prior_log_var_%d" % i, log_var_p)
tf.summary.scalar("kl_raw", tf.reduce_mean(kl_loss))
beta = self.get_beta(kl_loss)
# information capacity from "Understanding disentangling in beta-VAE"
if self.hparams.information_capacity > 0.0:
kl_loss = tf.abs(kl_loss - self.hparams.information_capacity)
return beta * kl_loss | [
"Get KL loss for all the predicted Gaussians."
] |
Please provide a description of the function:def construct_latent_tower(self, images, time_axis):
# No latent in the first phase
first_phase = tf.less(
self.get_iteration_num(), self.hparams.num_iterations_1st_stage)
# use all frames by default but this allows more
# predicted frames at inference time
latent_num_frames = self.hparams.latent_num_frames
tf.logging.info("Creating latent tower with %d frames." % latent_num_frames)
if latent_num_frames > 0:
images = images[:, :latent_num_frames]
return common_video.conv_latent_tower(
images=images,
time_axis=time_axis,
latent_channels=self.hparams.latent_channels,
min_logvar=self.hparams.latent_std_min,
is_training=self.is_training,
random_latent=first_phase,
tiny_mode=self.hparams.tiny_mode,
small_mode=self.hparams.small_mode) | [
"Create the latent tower."
] |
Please provide a description of the function:def transformer_encode(encoder_function, inputs, target_space, hparams,
attention_weights=None, features=None, losses=None,
**kwargs):
inputs = common_layers.flatten4d3d(inputs)
encoder_input, self_attention_bias, encoder_decoder_attention_bias = (
transformer_prepare_encoder(
inputs, target_space, hparams, features=features))
mlperf_log.transformer_print(
key=mlperf_log.MODEL_HP_LAYER_POSTPROCESS_DROPOUT,
value=hparams.layer_prepostprocess_dropout,
hparams=hparams)
encoder_input = tf.nn.dropout(encoder_input,
1.0 - hparams.layer_prepostprocess_dropout)
attn_bias_for_padding = None
# Otherwise the encoder will just use encoder_self_attention_bias.
if hparams.unidirectional_encoder:
attn_bias_for_padding = encoder_decoder_attention_bias
encoder_output = encoder_function(
encoder_input,
self_attention_bias,
hparams,
nonpadding=features_to_nonpadding(features, "inputs"),
save_weights_to=attention_weights,
make_image_summary=not common_layers.is_xla_compiled(),
losses=losses,
attn_bias_for_padding=attn_bias_for_padding,
**kwargs)
return encoder_output, encoder_decoder_attention_bias | [
"Encode transformer inputs.\n\n Args:\n encoder_function: the encoder function\n inputs: Transformer inputs [batch_size, input_length, 1, hidden_dim] which\n will be flattened along the two spatial dimensions.\n target_space: scalar, target space ID.\n hparams: hyperparameters for model.\n attention_weights: weight to store attention to.\n features: optionally pass the entire features dictionary as well. This is\n needed now for \"packed\" datasets.\n losses: optional list onto which to append extra training losses\n **kwargs: additional arguments to pass to encoder_function\n\n Returns:\n Tuple of:\n encoder_output: Encoder representation.\n [batch_size, input_length, hidden_dim]\n encoder_decoder_attention_bias: Bias and mask weights for\n encoder-decoder attention. [batch_size, input_length]\n "
] |
Please provide a description of the function:def transformer_decode(decoder_function,
decoder_input,
encoder_output,
encoder_decoder_attention_bias,
decoder_self_attention_bias,
hparams,
attention_weights=None,
cache=None,
decode_loop_step=None,
nonpadding=None,
losses=None,
**kwargs):
mlperf_log.transformer_print(
key=mlperf_log.MODEL_HP_LAYER_POSTPROCESS_DROPOUT,
value=hparams.layer_prepostprocess_dropout,
hparams=hparams)
decoder_input = tf.nn.dropout(decoder_input,
1.0 - hparams.layer_prepostprocess_dropout)
decoder_output = decoder_function(
decoder_input,
encoder_output,
decoder_self_attention_bias,
encoder_decoder_attention_bias,
hparams,
cache=cache,
decode_loop_step=decode_loop_step,
nonpadding=nonpadding,
save_weights_to=attention_weights,
losses=losses,
**kwargs)
if (common_layers.is_xla_compiled() and
hparams.mode == tf.estimator.ModeKeys.TRAIN):
# TPU does not react kindly to extra dimensions.
# TODO(noam): remove this once TPU is more forgiving of extra dims.
return decoder_output
else:
# Expand since t2t expects 4d tensors.
return tf.expand_dims(decoder_output, axis=2) | [
"Decode Transformer outputs from encoder representation.\n\n Args:\n decoder_function: the decoder function\n decoder_input: inputs to bottom of the model. [batch_size, decoder_length,\n hidden_dim]\n encoder_output: Encoder representation. [batch_size, input_length,\n hidden_dim]\n encoder_decoder_attention_bias: Bias and mask weights for encoder-decoder\n attention. [batch_size, input_length]\n decoder_self_attention_bias: Bias and mask weights for decoder\n self-attention. [batch_size, decoder_length]\n hparams: hyperparameters for model.\n attention_weights: weight to store attention to.\n cache: dict, containing tensors which are the results of previous\n attentions, used for fast decoding.\n decode_loop_step: An integer, step number of the decoding loop. Only used\n for inference on TPU.\n nonpadding: optional Tensor with shape [batch_size, decoder_length]\n losses: optional list onto which to append extra training losses\n **kwargs: additional arguments to pass to decoder_function\n\n Returns:\n Final decoder representation. [batch_size, decoder_length, hidden_dim]\n "
] |
Please provide a description of the function:def _init_transformer_cache(cache, hparams, batch_size, attention_init_length,
encoder_output, encoder_decoder_attention_bias,
scope_prefix):
key_channels = hparams.attention_key_channels or hparams.hidden_size
value_channels = hparams.attention_value_channels or hparams.hidden_size
num_layers = hparams.num_decoder_layers or hparams.num_hidden_layers
vars_3d_num_heads = (
hparams.num_heads if hparams.get("attention_variables_3d") else 0)
if cache is None:
cache = {}
cache.update({
"layer_%d" % layer: { # pylint: disable=g-complex-comprehension
"k":
common_attention.split_heads(
tf.zeros([batch_size,
attention_init_length,
key_channels]), hparams.num_heads),
"v":
common_attention.split_heads(
tf.zeros([batch_size,
attention_init_length,
value_channels]), hparams.num_heads),
} for layer in range(num_layers)
})
# If `ffn_layer` is in `["dense_relu_dense" or "conv_hidden_relu"]`, then the
# cache key "f" won't be used, which means that the` shape of cache["f"]`
# won't be changed to
# `[beamsize*batch_size, decode_length, hparams.hidden_size]` and may cause
# error when applying `nest.map reshape function` on it.
if hparams.ffn_layer not in ["dense_relu_dense", "conv_hidden_relu"]:
for layer in range(num_layers):
cache["layer_%d" % layer]["f"] = tf.zeros(
[batch_size, 0, hparams.hidden_size])
if encoder_output is not None:
for layer in range(num_layers):
layer_name = "layer_%d" % layer
with tf.variable_scope(
"%sdecoder/%s/encdec_attention/multihead_attention" %
(scope_prefix, layer_name)):
k_encdec = common_attention.compute_attention_component(
encoder_output,
key_channels,
name="k",
vars_3d_num_heads=vars_3d_num_heads)
k_encdec = common_attention.split_heads(k_encdec, hparams.num_heads)
v_encdec = common_attention.compute_attention_component(
encoder_output,
value_channels,
name="v",
vars_3d_num_heads=vars_3d_num_heads)
v_encdec = common_attention.split_heads(v_encdec, hparams.num_heads)
cache[layer_name]["k_encdec"] = k_encdec
cache[layer_name]["v_encdec"] = v_encdec
cache["encoder_output"] = encoder_output
cache["encoder_decoder_attention_bias"] = encoder_decoder_attention_bias
return cache | [
"Create the initial cache for Transformer fast decoding."
] |
Please provide a description of the function:def fast_decode_tpu(encoder_output,
encoder_decoder_attention_bias,
symbols_to_logits_fn,
hparams,
decode_length,
vocab_size,
init_cache_fn=_init_transformer_cache,
beam_size=1,
top_beams=1,
alpha=1.0,
sos_id=0,
eos_id=beam_search.EOS_ID,
batch_size=None,
force_decode_length=False,
scope_prefix="body/",
use_top_k_with_unique=True):
if encoder_output is not None:
batch_size = common_layers.shape_list(encoder_output)[0]
cache = init_cache_fn(None, hparams, batch_size, decode_length,
encoder_output, encoder_decoder_attention_bias,
scope_prefix)
mlperf_log.transformer_print(
key=mlperf_log.MODEL_HP_SEQ_BEAM_SEARCH,
value={
"vocab_size": vocab_size,
"batch_size": batch_size,
"beam_size": beam_size,
"alpha": alpha,
"max_decode_length": decode_length
},
hparams=hparams)
if beam_size > 1: # Beam Search
initial_ids = sos_id * tf.ones([batch_size], dtype=tf.int32)
decoded_ids, scores, _ = beam_search.beam_search(
symbols_to_logits_fn,
initial_ids,
beam_size,
decode_length,
vocab_size,
alpha,
states=cache,
eos_id=eos_id,
stop_early=(top_beams == 1),
use_tpu=True,
use_top_k_with_unique=use_top_k_with_unique)
if top_beams == 1:
decoded_ids = decoded_ids[:, 0, 1:]
scores = scores[:, 0]
else:
decoded_ids = decoded_ids[:, :top_beams, 1:]
scores = scores[:, :top_beams]
else: # Greedy
def inner_loop(i, hit_eos, next_id, decoded_ids, cache, log_prob):
logits, cache = symbols_to_logits_fn(next_id, i, cache)
log_probs = common_layers.log_prob_from_logits(logits)
temperature = getattr(hparams, "sampling_temp", 0.0)
keep_top = getattr(hparams, "sampling_keep_top_k", -1)
if hparams.sampling_method == "argmax":
temperature = 0.0
next_id = common_layers.sample_with_temperature(
logits, temperature, keep_top)
hit_eos |= tf.equal(next_id, eos_id)
log_prob_indices = tf.stack([tf.range(tf.to_int64(batch_size)), next_id],
axis=1)
log_prob += tf.gather_nd(log_probs, log_prob_indices)
next_id = tf.expand_dims(next_id, axis=1)
decoded_ids = tf.transpose(decoded_ids)
decoded_ids = inplace_ops.alias_inplace_update(
decoded_ids, i, tf.squeeze(next_id, axis=1))
decoded_ids = tf.transpose(decoded_ids)
return i + 1, hit_eos, next_id, decoded_ids, cache, log_prob
def is_not_finished(i, hit_eos, *_):
finished = i >= decode_length
if not force_decode_length:
finished |= tf.reduce_all(hit_eos)
return tf.logical_not(finished)
decoded_ids = tf.zeros([batch_size, decode_length], dtype=tf.int64)
hit_eos = tf.fill([batch_size], False)
next_id = sos_id * tf.ones([batch_size, 1], dtype=tf.int64)
initial_log_prob = tf.zeros([batch_size], dtype=tf.float32)
def compute_cache_shape_invariants(tensor):
return tf.TensorShape(tensor.shape.as_list())
_, _, _, decoded_ids, _, log_prob = tf.while_loop(
is_not_finished,
inner_loop, [
tf.constant(0), hit_eos, next_id, decoded_ids, cache,
initial_log_prob
],
shape_invariants=[
tf.TensorShape([]),
tf.TensorShape([batch_size]),
tf.TensorShape([batch_size, 1]),
tf.TensorShape([batch_size, decode_length]),
nest.map_structure(compute_cache_shape_invariants, cache),
tf.TensorShape([batch_size]),
])
scores = log_prob
return {"outputs": decoded_ids, "scores": scores} | [
"Given encoder output and a symbols to logits function, does fast decoding.\n\n Implements both greedy and beam search decoding for TPU, uses beam search iff\n beam_size > 1, otherwise beam search related arguments are ignored.\n\n Args:\n encoder_output: A tensor, output from encoder.\n encoder_decoder_attention_bias: A tensor, bias for use in encoder-decoder\n attention.\n symbols_to_logits_fn: Incremental decoding, function mapping triple `(ids,\n step, cache)` to symbol logits.\n hparams: Run hyperparameters.\n decode_length: An integer, how many additional timesteps to decode.\n vocab_size: Output vocabulary size.\n init_cache_fn: Function that returns the initial cache dict.\n beam_size: An integer, number of beams.\n top_beams: An integer, how many of the beams to return.\n alpha: A float that controls the length penalty. Larger the alpha, stronger\n the preference for longer translations.\n sos_id: Start-of-sequence symbol.\n eos_id: End-of-sequence symbol.\n batch_size: An integer, must be passed if there is no input.\n force_decode_length: A bool, whether to force the full decode length, or if\n False, stop when all beams hit eos_id.\n scope_prefix: str, prefix for decoder layer variable scopes.\n use_top_k_with_unique: bool, whether to use a fast (but decreased precision)\n top_k during beam search.\n\n Returns:\n A dict of decoding results {\n \"outputs\": integer `Tensor` of decoded ids of shape\n [batch_size, <= decode_length] if top_beams == 1 or\n [batch_size, top_beams, <= decode_length] otherwise\n \"scores\": decoding log probs from the beam search,\n None if using greedy decoding (beam_size=1)\n }.\n\n Raises:\n NotImplementedError: If beam size > 1 with partial targets.\n ",
"One step of greedy decoding."
] |
Please provide a description of the function:def fast_decode(encoder_output,
encoder_decoder_attention_bias,
symbols_to_logits_fn,
hparams,
decode_length,
vocab_size,
init_cache_fn=_init_transformer_cache,
beam_size=1,
top_beams=1,
alpha=1.0,
sos_id=0,
eos_id=beam_search.EOS_ID,
batch_size=None,
force_decode_length=False,
scope_prefix="body/",
cache=None):
if encoder_output is not None:
batch_size = common_layers.shape_list(encoder_output)[0]
cache = init_cache_fn(
cache=cache,
hparams=hparams,
batch_size=batch_size,
attention_init_length=0,
encoder_output=encoder_output,
encoder_decoder_attention_bias=encoder_decoder_attention_bias,
scope_prefix=scope_prefix)
if beam_size > 1: # Beam Search
initial_ids = sos_id * tf.ones([batch_size], dtype=tf.int32)
decoded_ids, scores, cache = beam_search.beam_search(
symbols_to_logits_fn,
initial_ids,
beam_size,
decode_length,
vocab_size,
alpha,
states=cache,
eos_id=eos_id,
stop_early=(top_beams == 1))
if top_beams == 1:
decoded_ids = decoded_ids[:, 0, 1:]
scores = scores[:, 0]
else:
decoded_ids = decoded_ids[:, :top_beams, 1:]
scores = scores[:, :top_beams]
else: # Greedy
def inner_loop(i, hit_eos, next_id, decoded_ids, cache, log_prob):
logits, cache = symbols_to_logits_fn(next_id, i, cache)
log_probs = common_layers.log_prob_from_logits(logits)
temperature = getattr(hparams, "sampling_temp", 0.0)
keep_top = getattr(hparams, "sampling_keep_top_k", -1)
if hparams.sampling_method == "argmax":
temperature = 0.0
next_id = common_layers.sample_with_temperature(
logits, temperature, keep_top)
hit_eos |= tf.equal(next_id, eos_id)
log_prob_indices = tf.stack([tf.range(tf.to_int64(batch_size)), next_id],
axis=1)
log_prob += tf.gather_nd(log_probs, log_prob_indices)
next_id = tf.expand_dims(next_id, axis=1)
decoded_ids = tf.concat([decoded_ids, next_id], axis=1)
return i + 1, hit_eos, next_id, decoded_ids, cache, log_prob
def is_not_finished(i, hit_eos, *_):
finished = i >= decode_length
if not force_decode_length:
finished |= tf.reduce_all(hit_eos)
return tf.logical_not(finished)
decoded_ids = tf.zeros([batch_size, 0], dtype=tf.int64)
hit_eos = tf.fill([batch_size], False)
next_id = sos_id * tf.ones([batch_size, 1], dtype=tf.int64)
initial_log_prob = tf.zeros([batch_size], dtype=tf.float32)
_, _, _, decoded_ids, cache, log_prob = tf.while_loop(
is_not_finished,
inner_loop, [
tf.constant(0), hit_eos, next_id, decoded_ids, cache,
initial_log_prob
],
shape_invariants=[
tf.TensorShape([]),
tf.TensorShape([None]),
tf.TensorShape([None, None]),
tf.TensorShape([None, None]),
nest.map_structure(beam_search.get_state_shape_invariants, cache),
tf.TensorShape([None]),
])
scores = log_prob
return {"outputs": decoded_ids, "scores": scores, "cache": cache} | [
"Given encoder output and a symbols to logits function, does fast decoding.\n\n Implements both greedy and beam search decoding, uses beam search iff\n beam_size > 1, otherwise beam search related arguments are ignored.\n\n Args:\n encoder_output: Output from encoder.\n encoder_decoder_attention_bias: a bias tensor for use in encoder-decoder\n attention\n symbols_to_logits_fn: Incremental decoding; function mapping triple `(ids,\n step, cache)` to symbol logits.\n hparams: run hyperparameters\n decode_length: an integer. How many additional timesteps to decode.\n vocab_size: Output vocabulary size.\n init_cache_fn: Function that returns the initial cache dict.\n beam_size: number of beams.\n top_beams: an integer. How many of the beams to return.\n alpha: Float that controls the length penalty. larger the alpha, stronger\n the preference for longer translations.\n sos_id: End-of-sequence symbol in beam search.\n eos_id: End-of-sequence symbol in beam search.\n batch_size: an integer scalar - must be passed if there is no input\n force_decode_length: bool, whether to force the full decode length, or if\n False, stop when all beams hit eos_id.\n scope_prefix: str, prefix for decoder layer variable scopes.\n cache: cache dictionary for additional predictions.\n\n Returns:\n A dict of decoding results {\n \"outputs\": integer `Tensor` of decoded ids of shape\n [batch_size, <= decode_length] if top_beams == 1 or\n [batch_size, top_beams, <= decode_length] otherwise\n \"scores\": decoding log probs from the beam search,\n None if using greedy decoding (beam_size=1)\n }\n\n Raises:\n NotImplementedError: If beam size > 1 with partial targets.\n ",
"One step of greedy decoding."
] |
Please provide a description of the function:def transformer_prepare_decoder(targets, hparams, features=None):
if hparams.causal_decoder_self_attention:
# Causal attention.
if hparams.prepend_mode == "prepend_inputs_full_attention":
decoder_self_attention_bias = (
common_attention.attention_bias_prepend_inputs_full_attention(
common_attention.embedding_to_padding(targets)))
else:
decoder_self_attention_bias = (
common_attention.attention_bias_lower_triangle(
common_layers.shape_list(targets)[1]))
else:
# Full attention.
decoder_padding = common_attention.embedding_to_padding(targets)
decoder_self_attention_bias = (
common_attention.attention_bias_ignore_padding(decoder_padding))
if features and "targets_segmentation" in features:
# "Packed" dataset - keep the examples from seeing each other.
targets_segmentation = features["targets_segmentation"]
targets_position = features["targets_position"]
decoder_self_attention_bias += common_attention.attention_bias_same_segment(
targets_segmentation, targets_segmentation)
else:
targets_position = None
if hparams.proximity_bias:
decoder_self_attention_bias += common_attention.attention_bias_proximal(
common_layers.shape_list(targets)[1])
decoder_input = common_layers.shift_right_3d(targets)
if hparams.pos == "timing":
if targets_position is not None:
decoder_input = common_attention.add_timing_signal_1d_given_position(
decoder_input, targets_position)
else:
decoder_input = common_attention.add_timing_signal_1d(decoder_input)
elif hparams.pos == "emb":
decoder_input = common_attention.add_positional_embedding(
decoder_input, hparams.max_length, "targets_positional_embedding",
targets_position)
if hparams.activation_dtype == "bfloat16":
decoder_self_attention_bias = tf.cast(decoder_self_attention_bias,
tf.bfloat16)
return (decoder_input, decoder_self_attention_bias) | [
"Prepare one shard of the model for the decoder.\n\n Args:\n targets: a Tensor.\n hparams: run hyperparameters\n features: optionally pass the entire features dictionary as well. This is\n needed now for \"packed\" datasets.\n\n Returns:\n decoder_input: a Tensor, bottom of decoder stack\n decoder_self_attention_bias: a bias tensor for use in decoder self-attention\n "
] |
Please provide a description of the function:def transformer_decoder(decoder_input,
encoder_output,
decoder_self_attention_bias,
encoder_decoder_attention_bias,
hparams,
cache=None,
decode_loop_step=None,
name="decoder",
nonpadding=None,
save_weights_to=None,
make_image_summary=True,
losses=None,
layer_collection=None,
recurrent_memory_by_layer=None,
chunk_number=None,
):
x = decoder_input
attention_dropout_broadcast_dims = (
common_layers.comma_separated_string_to_integer_list(
getattr(hparams, "attention_dropout_broadcast_dims", "")))
mlperf_log.transformer_print(
key=mlperf_log.MODEL_HP_NUM_HIDDEN_LAYERS,
value=hparams.num_decoder_layers or hparams.num_hidden_layers,
hparams=hparams)
mlperf_log.transformer_print(
key=mlperf_log.MODEL_HP_ATTENTION_DROPOUT,
value=hparams.attention_dropout,
hparams=hparams)
mlperf_log.transformer_print(
key=mlperf_log.MODEL_HP_ATTENTION_DENSE,
value={
"use_bias": "false",
"num_heads": hparams.num_heads,
"hidden_size": hparams.hidden_size
},
hparams=hparams)
with tf.variable_scope(name):
for layer in range(hparams.num_decoder_layers or hparams.num_hidden_layers):
layer_name = "layer_%d" % layer
layer_cache = cache[layer_name] if cache is not None else None
if recurrent_memory_by_layer is not None:
recurrent_memory = recurrent_memory_by_layer[layer_name]
else:
recurrent_memory = None
if layer < hparams.get("num_area_layers", 0):
max_area_width = hparams.get("max_area_width", 1)
max_area_height = hparams.get("max_area_height", 1)
memory_height = hparams.get("max_area_height", 1)
else:
max_area_width = 1
max_area_height = 1
memory_height = 1
with tf.variable_scope(layer_name):
with tf.variable_scope("self_attention"):
y = common_attention.multihead_attention(
common_layers.layer_preprocess(
x, hparams, layer_collection=layer_collection),
None,
decoder_self_attention_bias,
hparams.attention_key_channels or hparams.hidden_size,
hparams.attention_value_channels or hparams.hidden_size,
hparams.hidden_size,
hparams.num_heads,
hparams.attention_dropout,
attention_type=hparams.self_attention_type,
max_relative_position=hparams.max_relative_position,
heads_share_relative_embedding=(
hparams.heads_share_relative_embedding),
add_relative_to_values=hparams.add_relative_to_values,
save_weights_to=save_weights_to,
cache=layer_cache,
make_image_summary=make_image_summary,
dropout_broadcast_dims=attention_dropout_broadcast_dims,
max_length=hparams.get("max_length"),
decode_loop_step=decode_loop_step,
vars_3d=hparams.get("attention_variables_3d"),
activation_dtype=hparams.get("activation_dtype", "float32"),
weight_dtype=hparams.get("weight_dtype", "float32"),
layer_collection=layer_collection,
recurrent_memory=recurrent_memory,
chunk_number=chunk_number,
hard_attention_k=hparams.get("hard_attention_k", 0),
max_area_width=max_area_width,
max_area_height=max_area_height,
memory_height=memory_height,
area_key_mode=hparams.get("area_key_mode", "none"),
area_value_mode=hparams.get("area_value_mode", "none"),
training=(hparams.get("mode", tf.estimator.ModeKeys.TRAIN)
== tf.estimator.ModeKeys.TRAIN))
x = common_layers.layer_postprocess(x, y, hparams)
if encoder_output is not None:
with tf.variable_scope("encdec_attention"):
y = common_attention.multihead_attention(
common_layers.layer_preprocess(
x, hparams, layer_collection=layer_collection),
encoder_output,
encoder_decoder_attention_bias,
hparams.attention_key_channels or hparams.hidden_size,
hparams.attention_value_channels or hparams.hidden_size,
hparams.hidden_size,
hparams.num_heads,
hparams.attention_dropout,
max_relative_position=hparams.max_relative_position,
heads_share_relative_embedding=(
hparams.heads_share_relative_embedding),
add_relative_to_values=hparams.add_relative_to_values,
save_weights_to=save_weights_to,
cache=layer_cache,
make_image_summary=make_image_summary,
dropout_broadcast_dims=attention_dropout_broadcast_dims,
max_length=hparams.get("max_length"),
vars_3d=hparams.get("attention_variables_3d"),
activation_dtype=hparams.get("activation_dtype", "float32"),
weight_dtype=hparams.get("weight_dtype", "float32"),
layer_collection=layer_collection,
hard_attention_k=hparams.get("hard_attention_k", 0),
max_area_width=max_area_width,
max_area_height=max_area_height,
memory_height=memory_height,
area_key_mode=hparams.get("area_key_mode", "none"),
area_value_mode=hparams.get("area_value_mode", "none"),
training=(hparams.get("mode", tf.estimator.ModeKeys.TRAIN)
== tf.estimator.ModeKeys.TRAIN))
x = common_layers.layer_postprocess(x, y, hparams)
with tf.variable_scope("ffn"):
y = transformer_ffn_layer(
common_layers.layer_preprocess(
x, hparams, layer_collection=layer_collection),
hparams,
conv_padding="LEFT",
nonpadding_mask=nonpadding,
losses=losses,
cache=layer_cache,
decode_loop_step=decode_loop_step,
layer_collection=layer_collection)
x = common_layers.layer_postprocess(x, y, hparams)
# if normalization is done in layer_preprocess, then it should also be done
# on the output, since the output can grow very large, being the sum of
# a whole stack of unnormalized layer outputs.
mlperf_log.transformer_print(
key=mlperf_log.MODEL_HP_NORM,
value={"hidden_size": hparams.hidden_size})
return common_layers.layer_preprocess(
x, hparams, layer_collection=layer_collection) | [
"A stack of transformer layers.\n\n Args:\n decoder_input: a Tensor\n encoder_output: a Tensor\n decoder_self_attention_bias: bias Tensor for self-attention (see\n common_attention.attention_bias())\n encoder_decoder_attention_bias: bias Tensor for encoder-decoder attention\n (see common_attention.attention_bias())\n hparams: hyperparameters for model\n cache: dict, containing tensors which are the results of previous\n attentions, used for fast decoding.\n decode_loop_step: An integer, step number of the decoding loop. Only used\n for inference on TPU.\n name: a string\n nonpadding: optional Tensor with shape [batch_size, encoder_length]\n indicating what positions are not padding. This is used to mask out\n padding in convolutional layers. We generally only need this mask for\n \"packed\" datasets, because for ordinary datasets, no padding is ever\n followed by nonpadding.\n save_weights_to: an optional dictionary to capture attention weights for\n visualization; the weights tensor will be appended there under a string\n key created from the variable scope (including name).\n make_image_summary: Whether to make an attention image summary.\n losses: optional list onto which to append extra training losses\n layer_collection: A tensorflow_kfac.LayerCollection. Only used by the\n KFAC optimizer. Default is None.\n recurrent_memory_by_layer: Optional dict, mapping layer names to instances\n of transformer_memory.RecurrentMemory. Default is None.\n chunk_number: an optional integer Tensor with shape [batch] used to operate\n the recurrent_memory.\n\n Returns:\n y: a Tensors\n "
] |
Please provide a description of the function:def transformer_base_v1():
hparams = common_hparams.basic_params1()
hparams.norm_type = "layer"
hparams.hidden_size = 512
hparams.batch_size = 4096
hparams.max_length = 256
hparams.clip_grad_norm = 0. # i.e. no gradient clipping
hparams.optimizer_adam_epsilon = 1e-9
hparams.learning_rate_schedule = "legacy"
hparams.learning_rate_decay_scheme = "noam"
hparams.learning_rate = 0.1
hparams.learning_rate_warmup_steps = 4000
hparams.initializer_gain = 1.0
hparams.num_hidden_layers = 6
hparams.initializer = "uniform_unit_scaling"
hparams.weight_decay = 0.0
hparams.optimizer_adam_beta1 = 0.9
hparams.optimizer_adam_beta2 = 0.98
hparams.num_sampled_classes = 0
hparams.label_smoothing = 0.1
hparams.shared_embedding_and_softmax_weights = True
hparams.symbol_modality_num_shards = 16
# Add new ones like this.
hparams.add_hparam("filter_size", 2048)
# Layer-related flags. If zero, these fall back on hparams.num_hidden_layers.
hparams.add_hparam("num_encoder_layers", 0)
hparams.add_hparam("num_decoder_layers", 0)
# Attention-related flags.
hparams.add_hparam("num_heads", 8)
hparams.add_hparam("attention_key_channels", 0)
hparams.add_hparam("attention_value_channels", 0)
hparams.add_hparam("ffn_layer", "dense_relu_dense")
hparams.add_hparam("parameter_attention_key_channels", 0)
hparams.add_hparam("parameter_attention_value_channels", 0)
# All hyperparameters ending in "dropout" are automatically set to 0.0
# when not in training mode.
hparams.add_hparam("attention_dropout", 0.0)
hparams.add_hparam("attention_dropout_broadcast_dims", "")
hparams.add_hparam("relu_dropout", 0.0)
hparams.add_hparam("relu_dropout_broadcast_dims", "")
hparams.add_hparam("pos", "timing") # timing, none
hparams.add_hparam("nbr_decoder_problems", 1)
hparams.add_hparam("proximity_bias", False)
hparams.add_hparam("causal_decoder_self_attention", True)
hparams.add_hparam("use_pad_remover", True)
hparams.add_hparam("self_attention_type", "dot_product")
hparams.add_hparam("conv_first_kernel", 3)
hparams.add_hparam("attention_variables_3d", False)
hparams.add_hparam("use_target_space_embedding", True)
# These parameters are only used when ffn_layer=="local_moe_tpu"
hparams.add_hparam("moe_overhead_train", 1.0)
hparams.add_hparam("moe_overhead_eval", 2.0)
hparams.moe_num_experts = 16
hparams.moe_loss_coef = 1e-3
# If specified, use this value instead of problem name in metrics.py.
# This is useful for programs that can automatically compare experiments side
# by side based on the same metric names.
hparams.add_hparam("overload_eval_metric_name", "")
# For making a transformer encoder unidirectional by using masked
# attention.
hparams.add_hparam("unidirectional_encoder", False)
# For hard attention.
hparams.add_hparam("hard_attention_k", 0)
return hparams | [
"Set of hyperparameters."
] |
Please provide a description of the function:def transformer_base_v2():
hparams = transformer_base_v1()
hparams.layer_preprocess_sequence = "n"
hparams.layer_postprocess_sequence = "da"
hparams.layer_prepostprocess_dropout = 0.1
hparams.attention_dropout = 0.1
hparams.relu_dropout = 0.1
hparams.learning_rate_warmup_steps = 8000
hparams.learning_rate = 0.2
return hparams | [
"Set of hyperparameters."
] |
Please provide a description of the function:def transformer_base_vq_ada_32ex_packed():
hparams = transformer_base_v2()
expert_utils.update_hparams_for_vq_gating(hparams)
hparams.moe_num_experts = 32
hparams.gating_type = "vq"
# this gives us a batch size of 16 because each seq is len 256
hparams.batch_size = 5072
hparams.ffn_layer = "local_moe"
hparams.shared_embedding_and_softmax_weights = False
hparams.learning_rate_warmup_steps = 10000
# one epoch for languagemodel_lm1b32k_packed = 27200 steps w/ bsize 128
hparams.learning_rate_decay_steps = 27200
hparams.num_heads = 4
hparams.num_blocks = 1
hparams.moe_k = 1
hparams.num_decoder_layers = 6
hparams.label_smoothing = 0.
hparams.layer_prepostprocess_dropout = 0.1
hparams.layer_postprocess_sequence = "dan"
hparams.layer_preprocess_sequence = "none"
hparams.weight_decay = 1e-06
hparams.attention_dropout = 0.1
hparams.optimizer = "Adafactor"
hparams.learning_rate_schedule = "linear_warmup*rsqrt_decay*linear_decay"
hparams.activation_dtype = "float32"
hparams.learning_rate = 0.1
hparams.learning_rate_constant = 1.0
return hparams | [
"Set of hyperparameters for lm1b packed following tpu params."
] |
Please provide a description of the function:def transformer_base_vq1_16_nb1_packed_nda_b01_scales():
hparams = transformer_base_vq_ada_32ex_packed()
hparams.use_scales = int(True)
hparams.moe_num_experts = 16
hparams.moe_k = 1
hparams.beta = 0.1
hparams.layer_preprocess_sequence = "n"
hparams.layer_postprocess_sequence = "da"
hparams.ema = False
return hparams | [
"Set of hyperparameters."
] |
Please provide a description of the function:def transformer_base_vq1_16_nb1_packed_dan_b01_scales():
hparams = transformer_base_vq_ada_32ex_packed()
hparams.use_scales = int(True)
hparams.moe_num_experts = 16
hparams.moe_k = 1
hparams.beta = 0.1
hparams.ema = False
return hparams | [
"Set of hyperparameters."
] |
Please provide a description of the function:def transformer_base_vq1_16_nb1_packed_nda_b01_scales_dialog():
hparams = transformer_base_vq1_16_nb1_packed_nda_b01_scales()
hparams.batch_size = 2048
hparams.max_length = 1024
hparams.filter_size = 3072
return hparams | [
"Set of hyperparameters."
] |
Please provide a description of the function:def transformer_ada_lmpackedbase_dialog():
hparams = transformer_base_vq_ada_32ex_packed()
hparams.max_length = 1024
hparams.ffn_layer = "dense_relu_dense"
hparams.batch_size = 4096
return hparams | [
"Set of hyperparameters."
] |
Please provide a description of the function:def transformer_base_v3():
# Update parameters here, then occasionally cut a versioned set, e.g.
# transformer_base_v2.
hparams = transformer_base_v2()
hparams.optimizer_adam_beta2 = 0.997
# New way of specifying learning rate schedule.
# Equivalent to previous version.
hparams.learning_rate_schedule = (
"constant*linear_warmup*rsqrt_decay*rsqrt_hidden_size")
hparams.learning_rate_constant = 2.0
return hparams | [
"Base parameters for Transformer model."
] |
Please provide a description of the function:def transformer_big():
hparams = transformer_base()
hparams.hidden_size = 1024
hparams.filter_size = 4096
# Reduce batch size to 2048 from 4096 to be able to train the model on a GPU
# with 12 GB memory. For example, NVIDIA TITAN V GPU.
hparams.batch_size = 2048
hparams.num_heads = 16
hparams.layer_prepostprocess_dropout = 0.3
return hparams | [
"HParams for transformer big model on WMT."
] |
Please provide a description of the function:def transformer_tall():
hparams = transformer_base()
hparams.batch_size = 2048
hparams.hidden_size = 768
hparams.filter_size = 3072
hparams.num_hidden_layers = 12
hparams.num_heads = 12
hparams.label_smoothing = 0.0
hparams.max_length = 1024
hparams.eval_drop_long_sequences = True
hparams.multiproblem_mixing_schedule = "pretrain"
hparams.multiproblem_vocab_size = 65536
hparams.clip_grad_norm = 1.0
return hparams | [
"Hparams for transformer on LM for pretraining/finetuning/mixing."
] |
Please provide a description of the function:def transformer_tall_finetune_tied():
hparams = transformer_tall()
hparams.multiproblem_max_input_length = 750
hparams.multiproblem_max_target_length = 100
hparams.multiproblem_schedule_max_examples = 0
hparams.learning_rate_schedule = ("linear_warmup*constant*cosdecay")
hparams.learning_rate_constant = 5e-5
hparams.learning_rate_warmup_steps = 100
# Set train steps to learning_rate_decay_steps or less
hparams.learning_rate_decay_steps = 80000
hparams.multiproblem_target_eval_only = True
hparams.multiproblem_reweight_label_loss = True
hparams.multiproblem_label_weight = 1.0
hparams.optimizer = "true_adam"
return hparams | [
"Tied means fine-tune CNN/DM summarization as LM."
] |
Please provide a description of the function:def transformer_tall_finetune_uniencdec():
hparams = transformer_tall()
hparams.max_input_seq_length = 750
hparams.max_target_seq_length = 100
hparams.optimizer = "true_adam"
hparams.learning_rate_schedule = ("linear_warmup*constant*cosdecay")
hparams.learning_rate_decay_steps = 80000
hparams.learning_rate_constant = 5e-5
hparams.learning_rate_warmup_steps = 100
hparams.unidirectional_encoder = True
return hparams | [
"Fine-tune CNN/DM with a unidirectional encoder and decoder."
] |
Please provide a description of the function:def transformer_tall_train_uniencdec():
hparams = transformer_tall()
hparams.max_input_seq_length = 750
hparams.max_target_seq_length = 100
hparams.optimizer = "true_adam"
hparams.learning_rate_schedule = ("linear_warmup*constant*cosdecay")
hparams.learning_rate_decay_steps = 150000
hparams.learning_rate_constant = 2e-4
hparams.unidirectional_encoder = True
return hparams | [
"Train CNN/DM with a unidirectional encoder and decoder."
] |
Please provide a description of the function:def transformer_tall_finetune_textclass():
hparams = transformer_tall()
hparams.learning_rate_constant = 6.25e-5
hparams.learning_rate_schedule = ("linear_warmup*constant*linear_decay")
hparams.multiproblem_schedule_max_examples = 0
hparams.multiproblem_target_eval_only = True
hparams.learning_rate_warmup_steps = 50
# Set train steps to learning_rate_decay_steps or less
hparams.learning_rate_decay_steps = 25000
hparams.multiproblem_reweight_label_loss = True
hparams.multiproblem_label_weight = 0.95
return hparams | [
"Hparams for transformer on LM for finetuning on text class problems."
] |
Please provide a description of the function:def transformer_tall_pretrain_lm():
hparams = transformer_tall()
hparams.learning_rate_constant = 2e-4
hparams.learning_rate_schedule = ("linear_warmup*constant*cosdecay")
hparams.optimizer = "adam_w"
hparams.optimizer_adam_beta1 = 0.9
hparams.optimizer_adam_beta2 = 0.999
hparams.optimizer_adam_epsilon = 1e-8
# Set max examples to something big when pretraining only the LM, definitely
# something an order of magnitude bigger than number of train steps.
hparams.multiproblem_schedule_max_examples = 5e8
# Set train steps to learning_rate_decay_steps or less
hparams.learning_rate_decay_steps = 5000000
return hparams | [
"Hparams for transformer on LM pretraining (with 64k vocab)."
] |
Please provide a description of the function:def transformer_tall_pretrain_lm_tpu_adafactor():
hparams = transformer_tall_pretrain_lm()
update_hparams_for_tpu(hparams)
hparams.max_length = 1024
# For multi-problem on TPU we need it in absolute examples.
hparams.batch_size = 8
hparams.multiproblem_vocab_size = 2**16
return hparams | [
"Hparams for transformer on LM pretraining (with 64k vocab) on TPU."
] |
Please provide a description of the function:def transformer_tall_pretrain_lm_tpu_adafactor_large():
hparams = transformer_tall_pretrain_lm_tpu_adafactor()
hparams.hidden_size = 1024
hparams.num_heads = 16
hparams.filter_size = 32768 # max fitting in 16G memory is 49152, batch 2
hparams.batch_size = 4
hparams.multiproblem_mixing_schedule = "constant"
# Task order: lm/en-de/en-fr/en-ro/de-en/fr-en/ro-en/cnndm/mnli/squad.
hparams.multiproblem_per_task_threshold = "320,80,160,1,80,160,2,20,10,5"
return hparams | [
"Hparams for transformer on LM pretraining on TPU, large model."
] |
Please provide a description of the function:def transformer_tall_pretrain_lm_tpu():
hparams = transformer_tall_pretrain_lm_tpu_adafactor()
# Optimizer gets reset in update_hparams_for_tpu so we set it again here.
hparams.learning_rate_constant = 2e-4
hparams.learning_rate_schedule = ("linear_warmup * constant * cosdecay")
hparams.optimizer = "adam_w"
return hparams | [
"Hparams for transformer on LM pretraining on TPU with AdamW."
] |
Please provide a description of the function:def transformer_base_single_gpu():
hparams = transformer_base()
hparams.batch_size = 1024
hparams.learning_rate_schedule = "constant*linear_warmup*rsqrt_decay"
hparams.learning_rate_constant = 0.1
hparams.learning_rate_warmup_steps = 16000
return hparams | [
"HParams for transformer base model for single GPU."
] |
Please provide a description of the function:def transformer_parsing_base():
hparams = transformer_base()
hparams.attention_dropout = 0.2
hparams.layer_prepostprocess_dropout = 0.2
hparams.max_length = 512
hparams.learning_rate_warmup_steps = 16000
hparams.hidden_size = 1024
hparams.learning_rate = 0.05
hparams.shared_embedding_and_softmax_weights = False
return hparams | [
"HParams for parsing on WSJ only."
] |
Please provide a description of the function:def transformer_parsing_big():
hparams = transformer_big()
hparams.max_length = 512
hparams.shared_source_target_embedding = False
hparams.learning_rate_warmup_steps = 4000
hparams.layer_prepostprocess_dropout = 0.1
hparams.batch_size = 2048
hparams.learning_rate = 0.05
return hparams | [
"HParams for parsing on WSJ semi-supervised."
] |
Please provide a description of the function:def transformer_base_range(rhp):
# After starting from base, set intervals for some parameters.
rhp.set_float("learning_rate", 0.3, 3.0, scale=rhp.LOG_SCALE)
rhp.set_discrete("learning_rate_warmup_steps",
[1000, 2000, 4000, 8000, 16000])
rhp.set_float("initializer_gain", 0.5, 2.0)
rhp.set_float("optimizer_adam_beta1", 0.85, 0.95)
rhp.set_float("optimizer_adam_beta2", 0.97, 0.99)
rhp.set_float("weight_decay", 0.0, 1e-4) | [
"Small range of hyperparameters."
] |
Please provide a description of the function:def transformer_relative():
hparams = transformer_base()
hparams.pos = None
hparams.self_attention_type = "dot_product_relative"
hparams.max_relative_position = 20
return hparams | [
"Use relative position embeddings instead of absolute position encodings."
] |
Please provide a description of the function:def transformer_mlperf_tpu():
hparams = transformer_base_v3()
hparams.mlperf_mode = True
hparams.symbol_modality_num_shards = 1
hparams.max_length = 256 # ignored when using "_packed" problems
hparams.batch_size = 2048 # per-chip batch size matches the reference model
hparams.hidden_size = 1024
hparams.filter_size = 4096
hparams.num_heads = 16
hparams.attention_dropout_broadcast_dims = "0,1" # batch, heads
hparams.relu_dropout_broadcast_dims = "1" # length
hparams.layer_prepostprocess_dropout_broadcast_dims = "1" # length
return hparams | [
"HParams for Transformer model on TPU for MLPerf on TPU 2x2."
] |
Please provide a description of the function:def update_hparams_for_tpu(hparams):
# Adafactor uses less memory than Adam.
# switch to Adafactor with its recommended learning rate scheme.
hparams.optimizer = "Adafactor"
hparams.learning_rate_schedule = "rsqrt_decay"
hparams.learning_rate_warmup_steps = 10000
# Avoid an expensive concat on TPU.
# >1 shards helps with faster parameter distribution on multi-GPU machines
hparams.symbol_modality_num_shards = 1
# Adaptive batch sizes and sequence lengths are not supported on TPU.
# Instead, every batch has the same sequence length and the same batch size.
# Longer sequences are dropped and shorter ones are padded.
#
# It is therefore suggested to use a problem where examples have been combined
# to a longer length, e.g. the "_packed" problems.
#
# For problems with variable sequence lengths, this parameter controls the
# maximum sequence length. Shorter sequences are dropped and longer ones
# are padded.
#
# For problems with fixed sequence lengths - e.g. the "_packed" problems,
# this hyperparameter is ignored.
hparams.max_length = 64
# TPUs have less memory than GPUs, so decrease the batch size
hparams.batch_size = 2048
# Using noise broadcast in the dropout layers saves memory during training.
hparams.attention_dropout_broadcast_dims = "0,1" # batch, heads
hparams.relu_dropout_broadcast_dims = "1" # length
hparams.layer_prepostprocess_dropout_broadcast_dims = "1" # length
return hparams | [
"Change hparams to be compatible with TPU training."
] |
Please provide a description of the function:def transformer_tpu_range(rhp):
# After starting from base, set intervals for some parameters.
rhp.set_float("learning_rate", 0.3, 3.0, scale=rhp.LOG_SCALE)
rhp.set_discrete("learning_rate_warmup_steps",
[1000, 2000, 4000, 8000, 16000])
rhp.set_float("initializer_gain", 0.5, 2.0)
rhp.set_float("optimizer_adam_beta1", 0.85, 0.95)
rhp.set_float("optimizer_adam_beta2", 0.97, 0.99)
rhp.set_float("weight_decay", 0.0, 2.0) | [
"Small range of hyperparameters."
] |
Please provide a description of the function:def transformer_clean():
hparams = transformer_base_v2()
hparams.label_smoothing = 0.0
hparams.layer_prepostprocess_dropout = 0.0
hparams.attention_dropout = 0.0
hparams.relu_dropout = 0.0
hparams.max_length = 0
return hparams | [
"No dropout, label smoothing, max_length."
] |
Please provide a description of the function:def transformer_lm_tpu_0():
hparams = transformer_clean_big()
update_hparams_for_tpu(hparams)
hparams.num_heads = 4 # Heads are expensive on TPUs.
hparams.batch_size = 4096
hparams.shared_embedding_and_softmax_weights = False
hparams.layer_prepostprocess_dropout = 0.1
return hparams | [
"HParams for training languagemodel_lm1b8k on tpu. 92M Params."
] |
Please provide a description of the function:def transformer_librispeech_v1():
hparams = transformer_base()
hparams.num_heads = 4
hparams.filter_size = 1024
hparams.hidden_size = 256
hparams.num_encoder_layers = 5
hparams.num_decoder_layers = 3
hparams.learning_rate = 0.15
hparams.batch_size = 6000000
librispeech.set_librispeech_length_hparams(hparams)
return hparams | [
"HParams for training ASR model on LibriSpeech V1."
] |
Please provide a description of the function:def transformer_librispeech_v2():
hparams = transformer_base()
hparams.max_length = 1240000
hparams.max_input_seq_length = 1550
hparams.max_target_seq_length = 350
hparams.batch_size = 16
hparams.num_decoder_layers = 4
hparams.num_encoder_layers = 6
hparams.hidden_size = 384
hparams.learning_rate = 0.15
hparams.daisy_chain_variables = False
hparams.filter_size = 1536
hparams.num_heads = 2
hparams.ffn_layer = "conv_relu_conv"
hparams.conv_first_kernel = 9
hparams.weight_decay = 0
hparams.layer_prepostprocess_dropout = 0.2
hparams.relu_dropout = 0.2
return hparams | [
"HParams for training ASR model on LibriSpeech V2."
] |
Please provide a description of the function:def transformer_librispeech_tpu_v1():
hparams = transformer_librispeech_v1()
update_hparams_for_tpu(hparams)
hparams.batch_size = 16
librispeech.set_librispeech_length_hparams(hparams)
return hparams | [
"HParams for training ASR model on Librispeech on TPU v1."
] |
Please provide a description of the function:def transformer_librispeech_tpu_v2():
hparams = transformer_librispeech_v2()
update_hparams_for_tpu(hparams)
hparams.batch_size = 16
librispeech.set_librispeech_length_hparams(hparams)
return hparams | [
"HParams for training ASR model on Librispeech on TPU v2."
] |
Please provide a description of the function:def transformer_tpu_1b():
hparams = transformer_tpu()
hparams.hidden_size = 2048
hparams.filter_size = 8192
hparams.num_hidden_layers = 8
# smaller batch size to avoid OOM
hparams.batch_size = 1024
hparams.activation_dtype = "bfloat16"
hparams.weight_dtype = "bfloat16"
# maximize number of parameters relative to computation by not sharing.
hparams.shared_embedding_and_softmax_weights = False
return hparams | [
"Hparams for machine translation with ~1.1B parameters."
] |
Please provide a description of the function:def transformer_wikitext103_l4k_v0():
hparams = transformer_big()
# Adafactor uses less memory than Adam.
# switch to Adafactor with its recommended learning rate scheme.
hparams.optimizer = "Adafactor"
hparams.learning_rate_schedule = "rsqrt_decay"
hparams.learning_rate_warmup_steps = 10000
hparams.num_heads = 4
hparams.max_length = 4096
hparams.batch_size = 4096
hparams.shared_embedding_and_softmax_weights = False
hparams.num_hidden_layers = 8
hparams.attention_dropout = 0.1
hparams.layer_prepostprocess_dropout = 0.2
hparams.relu_dropout = 0.1
hparams.label_smoothing = 0.0
# Using noise broadcast in the dropout layers saves memory during training.
hparams.attention_dropout_broadcast_dims = "0,1" # batch, heads
hparams.relu_dropout_broadcast_dims = "1" # length
hparams.layer_prepostprocess_dropout_broadcast_dims = "1" # length
# Avoid an expensive concat on TPU.
# >1 shards helps with faster parameter distribution on multi-GPU machines
hparams.symbol_modality_num_shards = 1
return hparams | [
"HParams for training languagemodel_wikitext103_l4k."
] |
Please provide a description of the function:def transformer_wikitext103_l4k_memory_v0():
hparams = transformer_wikitext103_l4k_v0()
hparams.split_targets_chunk_length = 64
hparams.split_targets_max_chunks = 64
hparams.split_targets_strided_training = True
hparams.add_hparam("memory_type", "transformer_xl")
# The hparams specify batch size *before* chunking, but we want to have a
# consistent 4K batch size *after* chunking to fully utilize the hardware.
target_tokens_per_batch = 4096
hparams.batch_size = int(target_tokens_per_batch * (
hparams.max_length / hparams.split_targets_chunk_length)) # 262144
hparams.pos = None
hparams.self_attention_type = "dot_product_relative"
hparams.max_relative_position = 2 * hparams.split_targets_chunk_length
hparams.add_hparam("unconditional", True)
hparams.add_hparam("recurrent_memory_batch_size", 0) # 0 = try to guess
# By default, cache one chunk only (like Transformer-XL)
hparams.add_hparam("num_memory_items", hparams.split_targets_chunk_length)
return hparams | [
"HParams for training languagemodel_wikitext103_l4k with memory."
] |
Please provide a description of the function:def transformer_wikitext103_l16k_memory_v0():
hparams = transformer_wikitext103_l4k_memory_v0()
hparams.max_length = 16384
hparams.split_targets_chunk_length = 64
hparams.split_targets_max_chunks = int(
hparams.max_length / hparams.split_targets_chunk_length)
# The hparams specify batch size *before* chunking, but we want to have a
# consistent 4K batch size *after* chunking to fully utilize the hardware.
target_tokens_per_batch = 4096
hparams.batch_size = int(target_tokens_per_batch * (
hparams.max_length / hparams.split_targets_chunk_length))
hparams.max_relative_position = 2 * hparams.split_targets_chunk_length
return hparams | [
"HParams for training languagemodel_wikitext103_l16k with memory."
] |
Please provide a description of the function:def transformer_cifar10_memory_v0():
hparams = transformer_wikitext103_l4k_memory_v0()
hparams.num_hidden_layers = 6
hparams.max_length = 32 * 32 * 3
hparams.split_targets_chunk_length = 64 * 3
hparams.split_targets_max_chunks = int(
hparams.max_length / hparams.split_targets_chunk_length)
hparams.num_memory_items = 128 * 3
# Since this is an image problem, batch size refers to examples (not tokens)
target_images_per_batch = 4
hparams.batch_size = int(target_images_per_batch * (
hparams.max_length / hparams.split_targets_chunk_length))
# The recurrent memory needs to know the actual batch size (in sequences)
hparams.recurrent_memory_batch_size = hparams.batch_size
hparams.max_relative_position = (
hparams.num_memory_items + hparams.split_targets_chunk_length)
return hparams | [
"HParams for training image_cifar10_plain_gen_flat_rev with memory."
] |
Please provide a description of the function:def transformer_imagenet64_memory_v0():
hparams = transformer_cifar10_memory_v0()
hparams.max_length = 64 * 64 * 3
hparams.split_targets_chunk_length = 64 * 3
hparams.split_targets_max_chunks = int(
hparams.max_length / hparams.split_targets_chunk_length)
hparams.num_memory_items = 128 * 3
# Since this is an image problem, batch size refers to examples (not tokens)
target_images_per_batch = 2
hparams.batch_size = int(target_images_per_batch * (
hparams.max_length / hparams.split_targets_chunk_length))
# The recurrent memory needs to know the actual batch size (in sequences)
hparams.recurrent_memory_batch_size = hparams.batch_size
hparams.max_relative_position = 3072
return hparams | [
"HParams for training image_imagenet64_gen_flat_rev with memory."
] |
Please provide a description of the function:def maybe_reshape_4d_to_3d(x):
x_shape = common_layers.shape_list(x)
is_4d = False
if len(x_shape) == 4:
x = tf.reshape(x, [x_shape[0], x_shape[1]*x_shape[2], x_shape[3]])
is_4d = True
return x, x_shape, is_4d | [
"Reshape input from 4D to 3D if necessary."
] |
Please provide a description of the function:def local_attention_2d(x, hparams, attention_type="local_attention_2d"):
# self-attention
with tf.variable_scope("local_2d_self_att"):
y = common_attention.multihead_attention_2d(
x,
None,
hparams.attention_key_channels or hparams.hidden_size,
hparams.attention_value_channels or hparams.hidden_size,
hparams.hidden_size,
hparams.num_heads,
attention_type=attention_type,
query_shape=hparams.query_shape,
memory_flange=hparams.memory_flange,
name="self_attention")
return y | [
"Local 2d, self attention layer."
] |
Please provide a description of the function:def local_within_block_attention(x,
self_attention_bias,
hparams,
attention_type="local_within_block_mask_right",
q_padding="VALID",
kv_padding="VALID"):
x_new, x_shape, is_4d = maybe_reshape_4d_to_3d(x)
with tf.variable_scope("local_within_block"):
y = common_attention.multihead_attention(
common_layers.layer_preprocess(x_new, hparams),
None,
self_attention_bias,
hparams.attention_key_channels or hparams.hidden_size,
hparams.attention_value_channels or hparams.hidden_size,
hparams.hidden_size,
hparams.num_heads,
hparams.attention_dropout,
attention_type=attention_type,
block_width=hparams.block_width,
block_length=hparams.block_length,
q_padding=q_padding,
kv_padding=kv_padding,
q_filter_width=hparams.q_filter_width,
kv_filter_width=hparams.kv_filter_width,
name="local_within_block")
if is_4d:
y = tf.reshape(y, x_shape)
return y | [
"Local within block self attention."
] |
Please provide a description of the function:def local_attention_1d(x,
hparams,
attention_type="local_unmasked",
q_padding="VALID",
kv_padding="VALID"):
# self-attention
x, x_shape, is_4d = maybe_reshape_4d_to_3d(x)
with tf.variable_scope("local_1d_self_att"):
y = common_attention.multihead_attention(
x,
None,
None,
hparams.attention_key_channels or hparams.hidden_size,
hparams.attention_value_channels or hparams.hidden_size,
hparams.hidden_size,
hparams.num_heads,
hparams.attention_dropout,
attention_type=attention_type,
shared_rel=hparams.shared_rel,
block_width=hparams.block_width,
block_length=hparams.block_length,
q_padding=q_padding,
kv_padding=kv_padding,
q_filter_width=hparams.q_filter_width,
kv_filter_width=hparams.kv_filter_width,
make_image_summary=False,
name="self_attention")
if is_4d:
y = tf.reshape(y, x_shape)
return y | [
"Local 1d self attention."
] |
Please provide a description of the function:def get_dilated_1d_attention_mask(
num_heads, block_size,
num_blocks, memory_size, gap_size,
name="dilated_mask"):
mask = np.ones((num_heads, block_size, 2*block_size), np.bool)
# now going over every row to do the right assignment of
# memory blocks
for i in range(block_size):
visible = 2*block_size - (block_size-i)
# You always attend to yourself, set the mask for that
mask[:, i, -(block_size - i)] = 0
# Maybe num_blocks can be automatically calculated?
for j in range(num_blocks):
for k in range(memory_size):
index = ((gap_size + memory_size)*j) + k
if index >= visible:
break
mask[:, i, -(index + block_size - i + 1)] = 0 # Verify
# adding a num blocks dimension
mask = np.expand_dims(mask, axis=1)
return tf.constant(mask, dtype=tf.int32, name=name) | [
"Dilated attention with a masking strategy."
] |
Please provide a description of the function:def dilated_attention_1d(x,
hparams,
attention_type="masked_dilated_1d",
q_padding="VALID",
kv_padding="VALID",
gap_size=2):
# self-attention
x, x_shape, is_4d = maybe_reshape_4d_to_3d(x)
with tf.variable_scope("masked_dilated_1d"):
y = common_attention.multihead_attention(
x,
None,
None,
hparams.attention_key_channels or hparams.hidden_size,
hparams.attention_value_channels or hparams.hidden_size,
hparams.hidden_size,
hparams.num_heads,
hparams.attention_dropout,
attention_type=attention_type,
block_width=hparams.block_width,
block_length=hparams.block_length,
q_padding=q_padding,
kv_padding=kv_padding,
q_filter_width=hparams.q_filter_width,
kv_filter_width=hparams.kv_filter_width,
gap_size=gap_size,
num_memory_blocks=hparams.num_memory_blocks,
name="self_attention")
if is_4d:
y = tf.reshape(y, x_shape)
y.set_shape([None, None, None, hparams.hidden_size])
return y | [
"Dilated 1d self attention."
] |
Please provide a description of the function:def local_global_attention(x,
self_attention_bias,
hparams,
q_padding="LEFT",
kv_padding="LEFT"):
with tf.variable_scope("self_local_global_att"):
[x_global, x_local] = tf.split(x, 2, axis=-1)
split_hidden_size = int(hparams.hidden_size / 2)
split_heads = int(hparams.num_heads / 2)
if self_attention_bias is not None:
self_attention_bias = get_self_attention_bias(x)
y_global = common_attention.multihead_attention(
x_global,
None,
self_attention_bias,
hparams.attention_key_channels or split_hidden_size,
hparams.attention_value_channels or split_hidden_size,
split_hidden_size,
split_heads,
hparams.attention_dropout,
q_filter_width=hparams.q_filter_width,
kv_filter_width=hparams.kv_filter_width,
q_padding=q_padding,
kv_padding=kv_padding,
name="global_self_att")
y_local = common_attention.multihead_attention(
x_local,
None,
None,
hparams.attention_key_channels or split_hidden_size,
hparams.attention_value_channels or split_hidden_size,
split_hidden_size,
split_heads,
hparams.attention_dropout,
attention_type="local_masked",
block_length=hparams.block_length,
block_width=hparams.block_width,
q_filter_width=hparams.q_filter_width,
kv_filter_width=hparams.kv_filter_width,
q_padding=q_padding,
kv_padding=kv_padding,
name="local_self_att")
y = tf.concat([y_global, y_local], axis=-1)
return y | [
"Local and global 1d self attention."
] |
Please provide a description of the function:def full_self_attention(x,
self_attention_bias,
hparams,
q_padding="LEFT",
kv_padding="LEFT"):
x, x_shape, is_4d = maybe_reshape_4d_to_3d(x)
if self_attention_bias is not None:
self_attention_bias = get_self_attention_bias(x)
with tf.variable_scope("self_att"):
y = common_attention.multihead_attention(
x,
None,
self_attention_bias,
hparams.attention_key_channels or hparams.hidden_size,
hparams.attention_value_channels or hparams.hidden_size,
hparams.hidden_size,
hparams.num_heads,
hparams.attention_dropout,
q_filter_width=hparams.q_filter_width,
kv_filter_width=hparams.kv_filter_width,
q_padding=q_padding,
kv_padding=kv_padding,
name="self_att")
if is_4d:
y = tf.reshape(y, [x_shape[0], x_shape[1], x_shape[2], x_shape[3]])
y.set_shape([None, None, None, hparams.hidden_size])
return y | [
"Full self-attention layer."
] |
Please provide a description of the function:def encdec_attention_1d(x,
encoder_output,
encoder_decoder_attention_bias,
hparams):
x, x_shape, is_4d = maybe_reshape_4d_to_3d(x)
encoder_output, _, _ = maybe_reshape_4d_to_3d(encoder_output)
with tf.variable_scope("encdec_attention"):
# Encoder Decoder attention
y = common_attention.multihead_attention(
x,
encoder_output,
encoder_decoder_attention_bias,
hparams.attention_key_channels or hparams.hidden_size,
hparams.attention_value_channels or hparams.hidden_size,
hparams.hidden_size,
hparams.num_heads,
hparams.attention_dropout,
name="encdec_attention")
if is_4d:
y = tf.reshape(y, x_shape)
y.set_shape([None, None, None, hparams.hidden_size])
return y | [
"Local 1d self attention."
] |
Please provide a description of the function:def transformer_decoder_layers(inputs,
encoder_output,
num_layers,
hparams,
self_attention_bias=None,
encoder_decoder_attention_bias=None,
attention_type=AttentionType.LOCAL_2D,
losses=None,
name="transformer"):
x = inputs
x = tf.nn.dropout(x, 1.0 - hparams.layer_prepostprocess_dropout)
if attention_type == AttentionType.DILATED:
assert len(hparams.gap_sizes) == num_layers
for layer in range(num_layers):
with tf.variable_scope("%s_layer_%d" % (name, layer)):
# self-attention + skip connections
if attention_type == AttentionType.LOCAL_2D:
y = local_attention_2d(common_layers.layer_preprocess(x, hparams),
hparams,
attention_type="masked_local_attention_2d")
elif attention_type == AttentionType.LOCAL_1D:
y = local_attention_1d(common_layers.layer_preprocess(x, hparams),
hparams,
attention_type="local_mask_right",
q_padding="LEFT", kv_padding="LEFT")
elif attention_type == AttentionType.RELATIVE_LOCAL_1D:
y = local_attention_1d(
common_layers.layer_preprocess(x, hparams),
hparams,
attention_type="local_relative_mask_right",
q_padding="LEFT",
kv_padding="LEFT")
elif attention_type == AttentionType.NON_CAUSAL_1D:
y = local_attention_1d(common_layers.layer_preprocess(x, hparams),
hparams,
attention_type="local_unmasked",
q_padding="VALID", kv_padding="VALID")
elif attention_type == AttentionType.LOCAL_BLOCK:
y = local_within_block_attention(
common_layers.layer_preprocess(x, hparams),
self_attention_bias, hparams,
attention_type="local_within_block_mask_right",
q_padding="LEFT", kv_padding="LEFT")
elif attention_type == AttentionType.GLOCAL:
y = local_global_attention(common_layers.layer_preprocess(x, hparams),
self_attention_bias, hparams,
q_padding="LEFT", kv_padding="LEFT")
elif attention_type == AttentionType.DILATED:
y = dilated_attention_1d(common_layers.layer_preprocess(x, hparams),
hparams, q_padding="LEFT",
kv_padding="LEFT",
gap_size=hparams.gap_sizes[layer])
elif attention_type == AttentionType.GLOBAL:
y = full_self_attention(common_layers.layer_preprocess(x, hparams),
self_attention_bias, hparams,
q_padding="LEFT", kv_padding="LEFT")
x = common_layers.layer_postprocess(x, y, hparams)
# enc-dec attention + skip connections
if encoder_output is not None:
y = encdec_attention_1d(common_layers.layer_preprocess(x, hparams),
encoder_output,
encoder_decoder_attention_bias,
hparams)
x = common_layers.layer_postprocess(x, y, hparams)
# feed-fwd layers + skip connections
y = ffn_layer(common_layers.layer_preprocess(x, hparams), hparams,
losses=losses)
x = common_layers.layer_postprocess(x, y, hparams)
return common_layers.layer_preprocess(x, hparams) | [
"Multi layer transformer."
] |
Please provide a description of the function:def transformer_encoder_layers(inputs,
num_layers,
hparams,
attention_type=AttentionType.GLOBAL,
self_attention_bias=None,
q_padding="VALID",
kv_padding="VALID",
name="transformer"):
x = inputs
x = tf.nn.dropout(x, 1.0 - hparams.layer_prepostprocess_dropout)
for layer in range(num_layers):
# attention layers + skip connections
with tf.variable_scope("%s_layer_%d" % (name, layer)):
if attention_type == AttentionType.LOCAL_2D:
y = local_attention_2d(common_layers.layer_preprocess(x, hparams),
hparams,
attention_type="local_attention_2d")
elif attention_type == AttentionType.LOCAL_1D:
y = local_attention_1d(common_layers.layer_preprocess(x, hparams),
hparams,
attention_type="local_unmasked",
q_padding=q_padding, kv_padding=kv_padding)
elif attention_type == AttentionType.GLOBAL:
y = full_self_attention(common_layers.layer_preprocess(x, hparams),
self_attention_bias, hparams,
q_padding=q_padding, kv_padding=kv_padding)
x = common_layers.layer_postprocess(x, y, hparams)
# feed-fwd layer + skip connections
y = ffn_layer(common_layers.layer_preprocess(x, hparams), hparams)
x = common_layers.layer_postprocess(x, y, hparams)
return common_layers.layer_preprocess(x, hparams) | [
"Multi layer transformer encoder."
] |
Please provide a description of the function:def ffn_layer(x, hparams, losses=None):
with tf.variable_scope("ffn"):
if hparams.ffn_layer == "none":
return x
if hparams.ffn_layer == "conv_hidden_relu":
y = common_layers.dense_relu_dense(
x,
hparams.filter_size,
hparams.hidden_size,
dropout=hparams.relu_dropout)
elif hparams.ffn_layer == "normed_conv_hidden_relu":
y = common_layers.normed_conv_hidden_relu(
x,
hparams.norm_type,
hparams.layer_norm_epsilon,
hparams.filter_size,
hparams.hidden_size,
dropout=hparams.relu_dropout,
norm_name="convnorm")
elif hparams.ffn_layer == "self_attention_ffn":
x_shape = tf.shape(x)
x = tf.reshape(x, [x_shape[0], -1, hparams.hidden_size])
y = common_attention.ffn_self_attention_layer(
x, hparams.filter_size, hparams.hidden_size, hparams.num_parts,
hparams.attention_dropout, hparams.share_kv)
y = tf.reshape(y, x_shape)
elif hparams.ffn_layer == "local_moe_tpu":
overhead = (hparams.moe_overhead_train
if hparams.mode == tf.estimator.ModeKeys.TRAIN
else hparams.moe_overhead_eval)
x, x_shape, is_4d = maybe_reshape_4d_to_3d(x)
y, loss = expert_utils.local_moe_tpu(
x, hparams.filter_size // 2,
hparams.hidden_size,
hparams.moe_num_experts, overhead=overhead,
loss_coef=hparams.moe_loss_coef)
if is_4d:
y = tf.reshape(y, x_shape)
if losses is None:
raise ValueError(
"transformer_ffn_layer with type local_moe_tpu must pass in "
"a losses list")
losses.append(loss)
else:
assert hparams.ffn_layer == "glu_ffn"
y = common_layers.gated_linear_unit_layer(x)
return y | [
"ffn layer transformer."
] |
Please provide a description of the function:def get_self_attention_bias(x):
x_shape = common_layers.shape_list(x)
self_attention_bias = common_attention.attention_bias_lower_triangle(
x_shape[1])
return self_attention_bias | [
"Creates masked self attention bias.\n\n Args:\n x: A tensor of shape [batch, length, depth]\n\n Returns:\n self_attention_bias: A tensor of shape [length, length, 1]\n "
] |
Please provide a description of the function:def postprocess_image(x, rows, cols, hparams):
batch = common_layers.shape_list(x)[0]
x = tf.reshape(x, [batch, rows, cols, hparams.hidden_size])
likelihood = getattr(hparams, "likelihood", DistributionType.CAT)
if likelihood == DistributionType.DMOL:
depth = hparams.num_mixtures * 10
targets = tf.layers.dense(x,
depth,
use_bias=False,
activation=None,
name="output_conv")
else:
depth = 256
targets = tf.layers.dense(x,
depth,
use_bias=True,
activation=None,
name="output_conv")
if (hparams.mode == tf.estimator.ModeKeys.PREDICT and
hparams.block_raster_scan):
y = targets
yshape = common_layers.shape_list(y)
block_length = hparams.query_shape[0]
block_width = hparams.query_shape[1]
# Break into block row wise.
y = tf.reshape(y,
[batch, yshape[1] // block_length, block_length,
yshape[2], depth])
yshape = common_layers.shape_list(y)
# Break into blocks width wise.
y_blocks = tf.reshape(y,
[batch, yshape[1], yshape[2],
yshape[3] // block_width, block_width, depth])
# Reshape targets as [batch, num_blocks_rows, num_block_cols, block_length,
# block_width, depth].
targets = tf.transpose(y_blocks, [0, 1, 3, 2, 4, 5])
return targets | [
"Postprocessing after decoding.\n\n Args:\n x: Tensor of shape [batch, ...], where ... can be any rank such that the\n number of elements in x is batch * rows * cols * hparams.hidden_size.\n rows: Integer representing number of rows in a 2-D data point.\n cols: Integer representing number of columns in a 2-D data point.\n hparams: HParams set.\n\n Returns:\n Tensor of shape [batch, rows, cols, depth], where depth is\n hparams.num_mixtures * 10 if hparams.likelihood is DMOL, otherwise 256. In\n the special case of inference and block raster scan order, it is a Tensor\n of shape [batch, num_blocks_rows, num_block_cols, block_length, block_width,\n depth].\n "
] |
Please provide a description of the function:def prepare_encoder(inputs, hparams, attention_type="local_1d"):
x = prepare_image(inputs, hparams, name="enc_channels")
# Add position signals.
x = add_pos_signals(x, hparams, "enc_pos")
x_shape = common_layers.shape_list(x)
if attention_type == "local_1d":
x = tf.reshape(x, [x_shape[0], x_shape[1]*x_shape[2], hparams.hidden_size])
x.set_shape([None, None, hparams.hidden_size])
elif attention_type == "local_2d":
x.set_shape([None, None, None, hparams.hidden_size])
return x | [
"Prepare encoder for images."
] |
Please provide a description of the function:def prepare_decoder(targets, hparams):
targets_shape = common_layers.shape_list(targets)
channels = hparams.num_channels
curr_infer_length = None
# during training, images are [batch, IMG_LEN, IMG_LEN, 3].
# At inference, they are [batch, curr_infer_length, 1, 1]
if hparams.mode == tf.estimator.ModeKeys.PREDICT:
curr_infer_length = targets_shape[1]
if hparams.block_raster_scan:
assert hparams.img_len*channels % hparams.query_shape[1] == 0
assert hparams.img_len % hparams.query_shape[0] == 0
total_block_width = hparams.img_len*channels
# Decoding is in block raster scan order. We divide the image into
# hparams.query_shape blocks and then decode each block in raster scan.
# To make that compatible with our inference pipeline, pad the target so
# that rows is a multiple of query_shape and columns is a multiple of
# hparams.img_len*channels
curr_infer_length = targets_shape[1]
block_padding_factor = total_block_width * hparams.query_shape[0]
targets = tf.pad(targets, [
[0, 0], [0, -curr_infer_length % block_padding_factor],
[0, 0], [0, 0]])
num_blocks = total_block_width // hparams.query_shape[1]
# Reshape the image to represent blocks
target_blocks = tf.reshape(
targets, [targets_shape[0], -1, num_blocks, hparams.query_shape[0],
hparams.query_shape[1]])
# Transpose to read the image in 2D fashion.
targets = tf.transpose(target_blocks, [0, 1, 3, 2, 4])
else:
# add padding to make sure the size of targets is a multiple of img_height
# times number of channels. This is needed for positional encodings and
# for doing the RGB lookup.
padding_factor = channels * hparams.img_len
targets = tf.pad(targets, [
[0, 0], [0, -curr_infer_length % padding_factor], [0, 0], [0, 0]])
targets = tf.reshape(targets,
[targets_shape[0], -1, hparams.img_len, channels])
# Preprocess image
x = prepare_image(targets, hparams, name="dec_channels")
x_shape = common_layers.shape_list(x)
if (hparams.dec_attention_type == AttentionType.LOCAL_2D or
hparams.dec_attention_type == AttentionType.LOCAL_BLOCK):
x = common_attention.right_shift_blockwise(x, hparams.query_shape)
x = add_pos_signals(x, hparams, "dec_pos")
else:
# Add position signals
x = tf.reshape(x, [targets_shape[0],
x_shape[1]*x_shape[2], hparams.hidden_size])
x = common_layers.shift_right_3d(x)
x = tf.reshape(x, [targets_shape[0],
x_shape[1], x_shape[2], hparams.hidden_size])
x = add_pos_signals(x, hparams, "dec_pos")
x = common_layers.cast_like(x, targets)
return x, x_shape[1], x_shape[2] | [
"Prepare decoder for images."
] |
Please provide a description of the function:def create_output(decoder_output, rows, cols, targets, hparams):
del targets # unused arg
decoded_image = postprocess_image(decoder_output, rows, cols, hparams)
batch = common_layers.shape_list(decoded_image)[0]
depth = common_layers.shape_list(decoded_image)[-1]
likelihood = getattr(hparams, "likelihood", DistributionType.CAT)
if hparams.mode == tf.estimator.ModeKeys.PREDICT:
y = tf.reshape(decoded_image, [batch, -1, 1, 1, depth])
output = y[:, :rows, :, :, :]
elif likelihood == DistributionType.CAT:
# Unpack the cols dimension of the Categorical.
channels = hparams.num_channels
output = tf.reshape(decoded_image,
[batch, rows, cols // channels, channels, depth])
else:
output = decoded_image
return output | [
"Creates output from decoder output and vars.\n\n Args:\n decoder_output: Tensor of shape [batch, ...], where ... can be any rank such\n that the number of elements is batch * rows * cols * hparams.hidden_size.\n rows: Integer representing number of rows in a 2-D data point.\n cols: Integer representing number of columns in a 2-D data point.\n targets: Tensor of shape [batch, hparams.img_len, hparams.img_len,\n hparams.num_channels].\n hparams: HParams set.\n\n Returns:\n Tensor of shape [batch, hparams.img_len, hparams.img_len,\n hparams.num_mixtures * 10] if hparams.likelihood is DMOL, otherwise\n [batch, hparams.img_len, hparams.img_len, hparams.num_channels, 256].\n In the special case of predict mode, it is a Tensor of rank 5.\n "
] |
Please provide a description of the function:def get_channel_embeddings(io_depth, targets, hidden_size, name="channel"):
targets_split = tf.split(targets, io_depth, axis=3)
rgb_embedding_var = tf.get_variable("rgb_target_emb_%s" % name,
[256 * io_depth, hidden_size])
rgb_embedding_var = tf.identity(rgb_embedding_var)
rgb_embedding_var *= float(hidden_size)**0.5
channel_target_embs = []
for i in range(io_depth):
# Adding the channel offsets to get the right embedding since the
# embedding tensor has shape 256 * io_depth, hidden_size
target_ids = tf.squeeze(targets_split[i], axis=3) + i * 256
target_embs = common_layers.gather(rgb_embedding_var, target_ids)
channel_target_embs.append(target_embs)
return tf.concat(channel_target_embs, axis=-1) | [
"Get separate embedding for each of the channels."
] |
Please provide a description of the function:def simulate(self, action):
with tf.name_scope("environment/simulate"):
if action.dtype in (tf.float16, tf.float32, tf.float64):
action = tf.check_numerics(action, "action")
def step(action):
step_response = self._batch_env.step(action)
# Current env doesn't return `info`, but EnvProblem does.
# TODO(afrozm): The proper way to do this is to make T2TGymEnv return
# an empty info return value.
if len(step_response) == 3:
(observ, reward, done) = step_response
else:
(observ, reward, done, _) = step_response
return (observ, reward.astype(np.float32), done)
observ, reward, done = tf.py_func(
step, [action],
[self.observ_dtype, tf.float32, tf.bool], name="step")
reward = tf.check_numerics(reward, "reward")
reward.set_shape((len(self),))
done.set_shape((len(self),))
with tf.control_dependencies([self._observ.assign(observ)]):
return tf.identity(reward), tf.identity(done) | [
"Step the batch of environments.\n\n The results of the step can be accessed from the variables defined below.\n\n Args:\n action: Tensor holding the batch of actions to apply.\n\n Returns:\n Operation.\n "
] |
Please provide a description of the function:def _reset_non_empty(self, indices):
observ = tf.py_func(
self._batch_env.reset, [indices], self.observ_dtype, name="reset")
observ.set_shape(indices.get_shape().concatenate(self.observ_shape))
with tf.control_dependencies([
tf.scatter_update(self._observ, indices, observ)]):
return tf.identity(observ) | [
"Reset the batch of environments.\n\n Args:\n indices: The batch indices of the environments to reset; defaults to all.\n\n Returns:\n Batch tensor of the new observations.\n "
] |
Please provide a description of the function:def include_revision(revision_num, skip_factor=1.1):
if skip_factor <= 1.0:
return True
return (int(math.log1p(revision_num) / math.log(skip_factor)) != int(
math.log(revision_num + 2.0) / math.log(skip_factor))) | [
"Decide whether to include a revision.\n\n If the number of revisions is large, we exclude some revisions to avoid\n a quadratic blowup in runtime, since the article is likely also large.\n\n We make the ratio between consecutive included revision numbers\n appproximately equal to \"factor\".\n\n Args:\n revision_num: an integer\n skip_factor: a floating point number >= 1.0\n\n Returns:\n a boolean\n "
] |
Please provide a description of the function:def file_page_generator(my_file, max_page_size=2**28):
page_start = " <page>\n"
page_end = " </page>\n"
chunk_size = max_page_size
page_start = " <page>\n"
page_end = " </page>\n"
leftovers = ""
while True:
chunk = my_file.read(chunk_size)
if not chunk:
break
chunk = leftovers + chunk
current_pos = 0
while True:
start_pos = chunk.find(page_start, current_pos)
if start_pos == -1:
break
end_pos = chunk.find(page_end, start_pos)
if end_pos == -1:
if len(chunk) - start_pos > max_page_size:
leftovers = ""
else:
leftovers = chunk[start_pos:]
break
raw_page = chunk[start_pos + len(page_start):end_pos]
if len(raw_page) < max_page_size:
ret = parse_page(raw_page)
if ret:
yield ret
current_pos = end_pos + len(page_end) | [
"Read wikipedia pages from a history dump.\n\n Since some pages can be terabytes in size (with all the revisions),\n we limit page size to max_page_size bytes.\n\n Args:\n my_file: an open file object.\n max_page_size: an integer\n\n Yields:\n strings\n "
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.