Hi,
I run the same OCR model (ch_ppocr_mobile_v2.0_rec) in snpe 1.60, but got right result in cpu runtime and wrong result in gpu runtime. I don't know why. Additionly, in gpu runtime, the results of fp16 and fp32 are different, but both are wrong.
Opinions expressed in the content posted here are the personal opinions of the original authors, and do not necessarily reflect those of Qualcomm Incorporated or its subsidiaries (“Qualcomm”). The content is provided for informational purposes only and is not meant to be an endorsement or representation by Qualcomm or any other party. This site may also provide links or references to non-Qualcomm sites and resources. Qualcomm makes no representations, warranties, or other commitments whatsoever about any non-Qualcomm sites or third-party resources that may be referenced, accessible from, or linked to this site.
Dear developer,
Thanks for your sharing.
Could you please share the issue to us through goole drive or other cloud storage for more deeply analysis?
In general, the runtime of cpu and gpu executed with float dlc. These results are same as your training golden output.
If something wrong, I believe error produced from your model
BR.
Wei