[SYSTEMDS-3545] img_cutout_linearized #1845
Conversation
|
Hi @kahvecioglu Welcome to open source, I can suggest looking at the PR #1838 |
|
Hi @kahvecioglu, How is this PR going, is there some uncommitted progress? |
bin/systemds
Outdated
| @@ -396,7 +396,7 @@ fi | |||
| JARNAME=$(basename "$SYSTEMDS_JAR_FILE") | |||
|
|
|||
| # relative path to jar file | |||
| SYSTEMDS_JAR_FILE=$(realpath --relative-to=. "$(dirname "$SYSTEMDS_JAR_FILE")")${DIR_SEP}${JARNAME} | |||
| SYSTEMDS_JAR_FILE=$(grealpath --relative-to=. "$(dirname "$SYSTEMDS_JAR_FILE")")${DIR_SEP}${JARNAME} | |||
There was a problem hiding this comment.
does this fix it for mac ?
There was a problem hiding this comment.
Yes, it does. Should I change it back in my next commit?
| # ------------------------------------------------------------------------------------------ | ||
|
|
||
|
|
||
| m_img_cutout_linearized = function(Matrix[Double] img_in, Integer x, Integer y, Integer width, Integer height, Double fill_value, Integer s_cols, Integer s_rows) return (Matrix[Double] img_out) { |
There was a problem hiding this comment.
i would like some newlines to break this line.
| # height Height of the rectangle (must be positive) | ||
| # fill_value The value to set for the rectangle | ||
| # s_cols Width of a single image | ||
| # s_rows Height of a single image |
| # | ||
| # OUTPUT: | ||
| # ------------------------------------------------------------------------------------------ | ||
| # img_out Output images as linearized 2D matrix with top left corner at [1, 1] |
There was a problem hiding this comment.
for not obvious reasons we need to spaces between the img_out and the beginning of the documentation
| import java.util.Random; | ||
| // if A should be generated one row at a time | ||
| //import java.util.stream.Stream; | ||
| //import java.util.stream.DoubleStream; |
There was a problem hiding this comment.
remove commented lines like this.
| int width = random.nextInt(s_cols - x) + 1; | ||
| int height = random.nextInt(s_rows - y) + 1; | ||
| int fill_color = random.nextInt(256); | ||
| int n_imgs = random.nextInt(100) + 1; |
There was a problem hiding this comment.
this random behavior is not good, since each test then behaves randomly and we cannot reproduce.
Use a parameterized class instead like in PR #1903
| * double[] row = Stream.of(matrix).flatMapToDouble(DoubleStream::of).toArray(); | ||
| * A[i] = row; | ||
| * } | ||
| */ |
There was a problem hiding this comment.
While generating the input, does 'sparsity' refer to the sparsity of a single row image or the entire matrix? I left that block in there because I wasn't sure about that.
There was a problem hiding this comment.
sparsity refer to the entire matrix generated, The generated matrix is trying to get the sparsity you request, but will most likely never hit it exactly.
|
@Baunsgaard The script and the test are working but the checks keep failing. (which is also the case for #1916 ) What should I do to resolve this issue? |
Sometimes some of the tests fail online, we rerun in such cases. You can always look into the ones that fail and try to decipher if it is related to your code. |
Cutout method for row linearized images. This functionality is being developed as part of a student project.