react-native-tensorflow-lite icon indicating copy to clipboard operation
react-native-tensorflow-lite copied to clipboard
verified publisher icon jazzystring1

Quantized Int Model

Open GantMan opened this issue 6 years ago • 4 comments

Looks like the quantized int model is 1/4th the size. image

I tried removing the *4 on the ByteBuffer, but then I get a new error that just says Error: null

Help appreciated. I can post the code, too, if you'd like.

GantMan avatar Jan 22 '19 06:01 GantMan

The quartered byteBuffer error for posterity.

image

GantMan avatar Jan 22 '19 06:01 GantMan

Actually, I haven't commited yet the latest version that supports quantized model. I'll commit it later or tomorrow because I'm busy right now.

jazzystring1 avatar Jan 22 '19 08:01 jazzystring1

Sounds good! Thanks for your work!

GantMan avatar Jan 22 '19 15:01 GantMan

its been a while you haven't fixed this issue. please fix it so that we can implement it in our apps

sandeepsharma27 avatar Mar 20 '19 09:03 sandeepsharma27