You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: Documentation/current_iteration.md
+13-1Lines changed: 13 additions & 1 deletion
Original file line number
Diff line number
Diff line change
@@ -199,6 +199,9 @@ We added support for dilation convolution on the GPU, exposed by BrainScript, C+
199
199
### Deterministic Pooling
200
200
Now call `cntk.debug.force_deterministic()` will make max and average pooling deterministic, this behavior depend on cuDNN version 6 or later.
201
201
202
+
### Add Crop Node to Python API
203
+
Motivation in order to support some image segmentation network, we added Crop node to C++ and Python API. Crop node crops its first input along spatial axes so that the result matches the spatial size of its second (reference) input. All non-spatial dimensions are unchanged. Crop offsets can be specified directly, or computed automatically, by traversing the network, and matching centers of receptive fields between activations in the two inputs
204
+
202
205
## Performance
203
206
### Intel MKL update to improve inference speed on CPU by around 2x on AlexNet
204
207
This work is rolled over to next release due to dependency on test infrastructure updates.
@@ -215,16 +218,25 @@ We added the image feature support for TensorBoard. Now CNTK users can use Tenso
215
218
### Acknowledgments
216
219
We thank the following community members for their contributions:+
217
220
221
+
*[@arturl](https://github.com/arturl)
222
+
*[@chentaMS](https://github.com/chentaMS)
218
223
*[@cijianzy](https://github.com/cijianzy)
224
+
*[@DaveyBiggers](https://github.com/DaveyBiggers)
219
225
*[@duli2012](https://github.com/duli2012)
226
+
*[@eldakms](https://github.com/eldakms)
220
227
*[@FDecaYed](https://github.com/FDecaYed)
228
+
*[@junjieqian](https://github.com/junjieqian)
221
229
*[@karolzak](https://github.com/karolzak)
222
230
*[@KGWANGMIN](https://github.com/KGWANGMIN)
231
+
*[@mnidza](https://github.com/mnidza)
223
232
*[@ottolu](https://github.com/ottolu)
233
+
*[@raaaar](https://github.com/raaaar)
224
234
*[@StillKeepTry](https://github.com/StillKeepTry)
225
235
*[@taehoonlee](https://github.com/taehoonlee)
236
+
*[@vmazalov](https://github.com/vmazalov)
237
+
226
238
227
239
We apologize for any community contributions we might have overlooked in these release notes.
228
240
229
241
## Others
230
-
### Continue work on [Deep Learning Explained](https://www.edx.org/course/deep-learning-explained-microsoft-dat236x) course on edX.
242
+
### Continue work on [Deep Learning Explained](https://www.edx.org/course/deep-learning-explained-microsoft-dat236x) course on edX.
0 commit comments