From bd0b3357dc8ed0f12768a03c103506c17792470b Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?P=C3=A1ll=20Haraldsson?= Date: Mon, 17 Apr 2017 15:08:13 +0000 Subject: [PATCH] You may want to change text on TensorFlow I just changed the link on "beam search" to the issue that is recently closed. See the comment I link to. --- README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/README.md b/README.md index 5f661cc..2e87a1a 100644 --- a/README.md +++ b/README.md @@ -49,7 +49,7 @@ In CNTK (as in TensorFlow and Theano), a network is specified as a symbolic grap **New models** Since TF uses symbolic graph of vector operations approach, specifying a new network is fairly easy. Although it doesn't support symbolic loop yet (at least not well tested/documented, as of 05/2016), RNNs can be made easy and efficient using the [bucketing trick](https://www.tensorflow.org/versions/r0.8/tutorials/seq2seq/index.html#bucketing-and-padding). -However, TF has a major weakness in terms of modeling flexibility. Every computational flow has be constructed as a static graph. That makes some computations difficult, such as [beam search](https://github.com/tensorflow/tensorflow/issues/654) (which is used frequently in sequence prediction tasks). +However, TF has a major weakness in terms of modeling flexibility. Every computational flow has be constructed as a static graph. That makes some computations difficult, such as [beam search](https://github.com/tensorflow/tensorflow/issues/654#issuecomment-289563944) (which is used frequently in sequence prediction tasks). #### Theano