From 339db86e2b70a1961e877c59f3a33fa015d5a7a8 Mon Sep 17 00:00:00 2001 From: "A. Unique TensorFlower" Date: Thu, 7 Jan 2016 13:26:53 -0800 Subject: [PATCH] Update generated Op docs. Change: 111631697 --- tensorflow/g3doc/api_docs/python/array_ops.md | 1 + tensorflow/g3doc/api_docs/python/image.md | 21 ++++++++++++------- tensorflow/g3doc/api_docs/python/math_ops.md | 4 ++-- tensorflow/g3doc/api_docs/python/train.md | 10 ++++++--- 4 files changed, 24 insertions(+), 12 deletions(-) diff --git a/tensorflow/g3doc/api_docs/python/array_ops.md b/tensorflow/g3doc/api_docs/python/array_ops.md index c8f30101a75..6f59722d012 100644 --- a/tensorflow/g3doc/api_docs/python/array_ops.md +++ b/tensorflow/g3doc/api_docs/python/array_ops.md @@ -727,6 +727,7 @@ output[3, 2:, :, ...] = input[3, 2:, :, ...] ``` In contrast, if: + ```prettyprint # Given this: batch_dim = 2 diff --git a/tensorflow/g3doc/api_docs/python/image.md b/tensorflow/g3doc/api_docs/python/image.md index 5d8fa7a4236..4b0a148a143 100644 --- a/tensorflow/g3doc/api_docs/python/image.md +++ b/tensorflow/g3doc/api_docs/python/image.md @@ -207,7 +207,7 @@ resized_image = tf.image.resize_images(image, 299, 299) - - - -### `tf.image.resize_images(images, new_height, new_width, method=0)` {#resize_images} +### `tf.image.resize_images(images, new_height, new_width, method=0, align_corners=False)` {#resize_images} Resize `images` to `new_width`, `new_height` using the specified `method`. @@ -233,6 +233,9 @@ the same as `new_width`, `new_height`. To avoid distortions see * `new_height`: integer. * `new_width`: integer. * `method`: ResizeMethod. Defaults to `ResizeMethod.BILINEAR`. +* `align_corners`: bool. If true, exactly align all 4 cornets of the input and + output. Defaults to `false`. Only implemented for bilinear + interpolation method so far. ##### Raises: @@ -298,7 +301,7 @@ Input images can be of different types but output images are always float. - - - -### `tf.image.resize_bilinear(images, size, name=None)` {#resize_bilinear} +### `tf.image.resize_bilinear(images, size, align_corners=None, name=None)` {#resize_bilinear} Resize `images` to `size` using bilinear interpolation. @@ -311,6 +314,10 @@ Input images can be of different types but output images are always float. 4-D with shape `[batch, height, width, channels]`. * `size`: A 1-D int32 Tensor of 2 elements: `new_height, new_width`. The new size for the images. +* `align_corners`: An optional `bool`. Defaults to `False`. + If true, rescale input by (new_height - 1) / (height - 1), which + exactly aligns the 4 corners of images and resized images. If false, rescale + by new_height / height. Treat similarly the width dimension. * `name`: A name for the operation (optional). ##### Returns: @@ -685,7 +692,7 @@ Example: # Decode an image and convert it to HSV. rgb_image = tf.decode_png(..., channels=3) rgb_image_float = tf.convert_image_dtype(rgb_image, tf.float32) -hsv_image = tf.hsv_to_rgb(rgb_image) +hsv_image = tf.rgb_to_hsv(rgb_image) ``` - - - @@ -788,7 +795,7 @@ Convert `image` to `dtype`, scaling its values if needed. Images that are represented using floating point values are expected to have values in the range [0,1). Image data stored in integer data types are -expected to have values in the range `[0,MAX]`, wbere `MAX` is the largest +expected to have values in the range `[0,MAX]`, where `MAX` is the largest positive representable number for the data type. This op converts between data types, scaling the values appropriately before @@ -921,7 +928,7 @@ channel and then adjusts each component `x` of each pixel to Adjust the contrast of an image by a random factor. -Equivalent to `adjust_constrast()` but uses a `contrast_factor` randomly +Equivalent to `adjust_contrast()` but uses a `contrast_factor` randomly picked in the interval `[lower, upper]`. ##### Args: @@ -1010,7 +1017,7 @@ picked in the interval `[-max_delta, max_delta]`. ### `tf.image.adjust_saturation(image, saturation_factor, name=None)` {#adjust_saturation} -Adjust staturation of an RGB image. +Adjust saturation of an RGB image. This is a convenience method that converts an RGB image to float representation, converts it to HSV, add an offset to the saturation channel, @@ -1073,7 +1080,7 @@ Linearly scales `image` to have zero mean and unit norm. This op computes `(x - mean) / adjusted_stddev`, where `mean` is the average of all values in image, and -`adjusted_stddev = max(stddev, 1.0/srqt(image.NumElements()))`. +`adjusted_stddev = max(stddev, 1.0/sqrt(image.NumElements()))`. `stddev` is the standard deviation of all values in `image`. It is capped away from zero to protect against division by 0 when handling uniform images. diff --git a/tensorflow/g3doc/api_docs/python/math_ops.md b/tensorflow/g3doc/api_docs/python/math_ops.md index 66a9c20e046..6101d111674 100644 --- a/tensorflow/g3doc/api_docs/python/math_ops.md +++ b/tensorflow/g3doc/api_docs/python/math_ops.md @@ -408,7 +408,7 @@ Computes exponential of x element-wise. \\(y = e^x\\). ### `tf.log(x, name=None)` {#log} -Computes natural logrithm of x element-wise. +Computes natural logarithm of x element-wise. I.e., \\(y = \log_e x\\). @@ -2018,7 +2018,7 @@ Computes the inverse permutation of a tensor. This operation computes the inverse of an index permutation. It takes a 1-D integer tensor `x`, which represents the indices of a zero-based array, and -swaps each value with its index position. In other words, for an ouput tensor +swaps each value with its index position. In other words, for an output tensor `y` and an input tensor `x`, this operation computes the following: `y[x[i]] = i for i in [0, 1, ..., len(x) - 1]` diff --git a/tensorflow/g3doc/api_docs/python/train.md b/tensorflow/g3doc/api_docs/python/train.md index 5423ba41cbd..7674e4fa725 100644 --- a/tensorflow/g3doc/api_docs/python/train.md +++ b/tensorflow/g3doc/api_docs/python/train.md @@ -298,6 +298,8 @@ Construct a new gradient descent optimizer. Optimizer that implements the Adagrad algorithm. +See http://www.jmlr.org/papers/volume12/duchi11a/duchi11a.pdf. + - - - #### `tf.train.AdagradOptimizer.__init__(learning_rate, initial_accumulator_value=0.1, use_locking=False, name='Adagrad')` {#AdagradOptimizer.__init__} @@ -350,14 +352,14 @@ Construct a new Momentum optimizer. Optimizer that implements the Adam algorithm. +See http://arxiv.org/pdf/1412.6980v7.pdf. + - - - #### `tf.train.AdamOptimizer.__init__(learning_rate=0.001, beta1=0.9, beta2=0.999, epsilon=1e-08, use_locking=False, name='Adam')` {#AdamOptimizer.__init__} Construct a new Adam optimizer. -Implementation is based on: http://arxiv.org/pdf/1412.6980v7.pdf - Initialization: ``` @@ -461,6 +463,8 @@ using this function. Optimizer that implements the RMSProp algorithm. +See http://www.cs.toronto.edu/~tijmen/csc321/slides/lecture_slides_lec6.pdf. + - - - #### `tf.train.RMSPropOptimizer.__init__(learning_rate, decay=0.9, momentum=0.0, epsilon=1e-10, use_locking=False, name='RMSProp')` {#RMSPropOptimizer.__init__} @@ -893,7 +897,7 @@ ema = tf.train.ExponentialMovingAverage(decay=0.9999) maintain_averages_op = ema.apply([var0, var1]) # Create an op that will update the moving averages after each training -# step. This is what we will use in place of the usuall trainig op. +# step. This is what we will use in place of the usual training op. with tf.control_dependencies([opt_op]): training_op = tf.group(maintain_averages_op)