onnxruntime
Speedup and reduce binary size for TfIdfVectorizer
#3197
Merged

Speedup and reduce binary size for TfIdfVectorizer #3197

yuslepukhin merged 13 commits into master from yuslepukhin/speedup_tfidf
yuslepukhin
yuslepukhin Speed up TfIdf history begins.
048949ba
yuslepukhin Use ParallelFor() for each of the rows processing.
02f52720
yuslepukhin Make it non-template, batch it.
3ea23146
yuslepukhin Parallellize processing.
d45af680
yuslepukhin Re-work. TODO: Tests failing.
a872cf59
yuslepukhin Handle UniGrams only once.
043963aa
yuslepukhin Batch all types.
c4774c4a
yuslepukhin Check for short tail within the inner loop.
e86a9d8a
yuslepukhin Merge branch 'master' into yuslepukhin/speedup_tfidf
5c97dcf8
yuslepukhin yuslepukhin requested a review from skottmckay skottmckay 5 years ago
yuslepukhin yuslepukhin requested a review from pranavsharma pranavsharma 5 years ago
yuslepukhin yuslepukhin requested a review 5 years ago
yuslepukhin yuslepukhin requested a review from hariharans29 hariharans29 5 years ago
yuslepukhin yuslepukhin requested a review from BowenBao BowenBao 5 years ago
skottmckay
skottmckay commented on 2020-03-12
yuslepukhin Rework recursive definition.
edcdf19c
skottmckay
skottmckay commented on 2020-03-12
skottmckay
skottmckay commented on 2020-03-12
yuslepukhin Use forward decl for recursive class definition.
1084ba93
yuslepukhin Make GCC happy.
2e88b9d2
yuslepukhin yuslepukhin requested a review from snnn snnn 5 years ago
snnn
snnn commented on 2020-03-13
yuslepukhin TryBatchParallelFor will do a good job computing the number of batches.
90911e4f
snnn
snnn approved these changes on 2020-03-13
yuslepukhin yuslepukhin merged 2a6e5ce9 into master 5 years ago
yuslepukhin yuslepukhin deleted the yuslepukhin/speedup_tfidf branch 5 years ago

Login to write a write a comment.

Login via GitHub

Assignees
No one assigned
Labels
Milestone