diff options
| author | Narumol Prangnawarat <narumol.prangnawarat@arm.com> | 2023-08-11 16:09:26 +0100 |
|---|---|---|
| committer | Narumol Prangnawarat <narumol.prangnawarat@arm.com> | 2023-08-21 13:34:54 +0100 |
| commit | 5f94124ac11afbbf2d2a4cda539b316964802c76 (patch) | |
| tree | 38d6f31517fe322cb0065a8c3e947801791a6405 /src/armnnTfLiteParser | |
| parent | c4f42340bd3d6664098c69d2fb044089aa39aea0 (diff) | |
| download | armnn-5f94124ac11afbbf2d2a4cda539b316964802c76.tar.gz | |
IVGCVSW-7964 Fix UnidirectionalSequenceLstm
* Fix incorrect batch size and time size
* Fix incorrect time major when max time =1
* Fix incorrect permutation
* Fix incorrect scratch buffer
* Unit tests
Signed-off-by: Narumol Prangnawarat <narumol.prangnawarat@arm.com>
Change-Id: I510fae55528be412a58d020e82bd283852e7800b
Diffstat (limited to 'src/armnnTfLiteParser')
| -rw-r--r-- | src/armnnTfLiteParser/TfLiteParser.cpp | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/src/armnnTfLiteParser/TfLiteParser.cpp b/src/armnnTfLiteParser/TfLiteParser.cpp index 301989ebb7..052aac6101 100644 --- a/src/armnnTfLiteParser/TfLiteParser.cpp +++ b/src/armnnTfLiteParser/TfLiteParser.cpp @@ -4049,7 +4049,7 @@ void TfLiteParserImpl::ParseUnidirectionalSequenceLSTM(size_t subgraphIndex, siz desc.m_HiddenStateScale = hiddentensor->GetInfo().GetQuantizationScale(); desc.m_HiddenStateZeroPoint = hiddentensor->GetInfo().GetQuantizationOffset(); } - unsigned int batchSize = inputTensorInfo.GetShape()[0]; + unsigned int batchSize = desc.m_TimeMajor ? inputTensorInfo.GetShape()[1] : inputTensorInfo.GetShape()[0]; unsigned int outputSize = outputTensorInfo.GetShape()[2]; unsigned int numUnits = cellStateInInfo.GetShape()[1]; |