* Allow int8 quantization for export_tfjs

--int8 param currently has no effect on export_tfjs. With this change, 
` python export.py --weights ../path/to/best.pt --include tfjs --int8` will add the --quantize_uint8 param to the tensorflowjs_converter script, greatly reducing model size for web usage.
Signed-off-by: David Strahm <david.strahm@lambda-it.ch>

* Update Dockerfile
Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com>

---------
Signed-off-by: David Strahm <david.strahm@lambda-it.ch>
Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com>
Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com>