在遵循 tensorflow 圖像分類教程時(shí)，首先它會(huì)緩存每個(gè)圖像的瓶頸：定義：cache_bottlenecks())我已經(jīng)使用 tensorflow 的Estimator. 這確實(shí)簡(jiǎn)化了所有代碼。但是我想在這里緩存瓶頸功能。這是我的model_fn. 我想緩存dense層的結(jié)果，這樣我就可以對(duì)實(shí)際訓(xùn)練進(jìn)行更改，而不必每次都計(jì)算瓶頸。我怎樣才能做到這一點(diǎn)？def model_fn(features, labels, mode, params): is_training = mode == tf.estimator.ModeKeys.TRAIN num_classes = len(params['label_vocab']) module = hub.Module(params['module_spec'], trainable=is_training and params['train_module']) bottleneck_tensor = module(features['image']) with tf.name_scope('final_retrain_ops'): logits = tf.layers.dense(bottleneck_tensor, units=num_classes, trainable=is_training) # save this? def train_op_fn(loss): optimizer = tf.train.AdamOptimizer() return optimizer.minimize(loss, global_step=tf.train.get_global_step()) head = tf.contrib.estimator.multi_class_head(n_classes=num_classes, label_vocabulary=params['label_vocab']) return head.create_estimator_spec( features, mode, logits, labels, train_op_fn=train_op_fn )
2 回答
慕仙森
TA貢獻(xiàn)1827條經(jīng)驗(yàn) 獲得超8個(gè)贊
TF 無(wú)法在您編碼時(shí)工作。你應(yīng)該：
從原始網(wǎng)絡(luò)導(dǎo)出瓶頸到文件。
使用瓶頸結(jié)果作為輸入，使用另一個(gè)網(wǎng)絡(luò)來(lái)訓(xùn)練您的數(shù)據(jù)。
守著一只汪
TA貢獻(xiàn)1872條經(jīng)驗(yàn) 獲得超4個(gè)贊
這樣的事情應(yīng)該工作（未經(jīng)測(cè)試）：
# Serialize the data into two tfrecord files
tf.enable_eager_execution()
feature_extractor = ...
features_file = tf.python_io.TFRecordWriter('features.tfrec')
label_file = tf.python_io.TFRecordWriter('labels.tfrec')
for images, labels in dataset:
features = feature_extractor(images)
features_file.write(tf.serialize_tensor(features))
label_file.write(tf.serialize_tensor(labels))
# Parse the files and zip them together
def parse(type, shape):
_def parse(x):
result = tf.parse_tensor(x, out_type=shape)
result = tf.reshape(result, FEATURE_SHAPE)
return result
return parse
features_ds = tf.data.TFRecordDataset('features.tfrec')
features_ds = features_ds.map(parse(tf.float32, FEATURE_SHAPE), num_parallel_calls=AUTOTUNE)
labels_ds = tf.data.TFRecordDataset('labels.tfrec')
labels_ds = labels_ds.map(parse(tf.float32, FEATURE_SHAPE), num_parallel_calls=AUTOTUNE)
ds = tf.data.Dataset.zip(features_ds, labels_ds)
ds = ds.unbatch().shuffle().repeat().batch().prefetch()...
您也可以使用 來(lái)完成它Dataset.cache，但我不是 100% 確定細(xì)節(jié)。
添加回答
舉報(bào)
0/150
提交
取消