TensorFlow v0.11.0rc0 發布,一個表達機器學習算法的接口

TensorFlow 是一個表達機器學習算法的接口，并且是執行算法的實現框架。使用 TensorFlow 表示的計算可以在眾多異構的系統上方便地移植，從移動設別如手機或者平板電腦到成千的GPU計算集群上都可以執行。該系統靈活，可以被用來表示很多的算法包括，深度神經網絡的訓練和推斷算法，也已經被用作科研和應用機器學習系統在若干的計算機科學領域或者其他領域中，例如語言識別、計算機視覺、機器人、信息檢索、自然語言理解、地理信息抽取和計算藥物發現。

更新日志

主要功能和改進

• cuDNN 5 support.
• HDFS Support.
• Adds Fused LSTM support via cuDNN 5 in tensorflow/contrib/cudnn_rnn.
• Improved support for NumPy style basic slicing including non-1 strides, ellipses, newaxis, and negative indices. For example complicated expressions like foo[1, 2:4, tf.newaxis, ..., :-3:-1, :] are now supported. In addition we have preliminary (non-broadcasting) support for sliced assignment to variables. In particular one can write var[1:3].assign([1,11,111]).
• Introducing core/util/tensor_bundle module: a module to efficiently serialize/deserialize tensors to disk. Will be used in TF's new checkpoint format.
• Added tf.svd for computing the singular value decomposition (SVD) of dense matrices or batches of matrices (CPU only).
• Added gradients for eigenvalues and eigenvectors computed using self_adjoint_eig orself_adjoint_eigvals.
• Eliminated batch_* methods for most linear algebra and FFT ops and promoted the non-batch version of the ops to handle batches of matrices.
• Tracing/timeline support for distributed runtime (no GPU profiler yet).
• C API gives access to inferred shapes with TF_GraphGetTensorNumDims and TF_GraphGetTensorShape.
• Shape functions for core ops have moved to C++ via REGISTER_OP(...).SetShapeFn(...). Python shape inference RegisterShape calls use the C++ shape functions withcommon_shapes.call_cpp_shape_fn. A future release will remove RegisterShape from python.

Bug修正等變化

• Documentation now includes operator overloads on Tensor and Variable.
• tensorflow.__git_version__ now allows users to identify the version of the code that TensorFlow was compiled with. We also have tensorflow.__git_compiler__ which identifies the compiler used to compile TensorFlow's core.
• Improved multi-threaded performance of batch_matmul.
• LSTMCell, BasicLSTMCell, and MultiRNNCell constructors now default to state_is_tuple=True. For a quick fix while transitioning to the new default, simply pass the argument state_is_tuple=False.
• DeviceFactory's AddDevices and CreateDevices functions now return a Status instead of void.
• Int32 elements of list(type) arguments are no longer placed in host memory by default. If necessary, a list(type) argument to a kernel can be placed in host memory using a HostMemory annotation.
• uniform_unit_scaling_initializer() no longer takes a full_shape arg, instead relying on the partition info passed to the initializer function when it's called.
• The NodeDef protocol message is now defined in its own file node_def.proto instead of graph.proto.
• ops.NoGradient was renamed ops.NotDifferentiable. ops.NoGradient will be removed soon.
• dot.h / DotGraph was removed (it was an early analysis tool prior to TensorBoard, no longer that useful). It remains in history should someone find the code useful.

下載

• Source code (zip)
• Source code (tar.gz)

本站原創，轉載時保留以下信息：
本文轉自：深度開源（open-open.com）
原文地址：http://www.baiduhome.net/news/view/16db7e2c