Skip to content

Commit 9c9cda0

Browse files
committed
update
1 parent b0c4d87 commit 9c9cda0

1 file changed

Lines changed: 7 additions & 6 deletions

File tree

README.md

Lines changed: 7 additions & 6 deletions
Original file line numberDiff line numberDiff line change
@@ -219,7 +219,7 @@ This model was trained for 8 epochs using the full text of The Time Machine nove
219219

220220
To verify some functions more quickly, I have introduced a handwritten digit recognition program.
221221

222-
`
222+
```
223223
./handwritten_recognition
224224
images magic : 2051
225225
label magic : 2049
@@ -230,22 +230,23 @@ Actions:
230230
evaluating : [10000/10000] correct : 9501
231231
epoch : 9 [50000/50000] loss : 0.150985
232232
evaluating : [10000/10000] correct : 9493
233-
`
233+
```
234234

235235
### graphviz supported
236236

237237
You can add a line of code like this to the program to output an out.dot file that records the tensor computation topology. For example, in mnist.cpp:
238-
`
238+
239+
```
239240
printAllActions();
240241
printDotGraph(); // here
241242
allocMemAndInitTensors();
242-
`
243+
```
243244

244245
If you have Graphviz installed, you can use the following command to convert the out.dot file into a PNG image:
245246

246-
`
247+
```
247248
dot -Tpng out.dot -o out.png
248-
`
249+
```
249250

250251
Here's an example from my side where a PNG file is generated as output.
251252

0 commit comments

Comments
 (0)