Mitchell Wortsman
418
posts
@Mitchnw
Joined October 2011
Pinned
Check out OpenFlamingo V2!
Excited it’s finally here — check out DataComp!
Sharing our project on 1) accelerating and 2) stabilizing training for large language-vision models 1) Towards accelerating training, we introduce SwitchBack, a linear layer for int8 quantized training which matches bfloat16 within 0.1 for CLIP ViT-Huge arxiv.org/abs/2304.13013
Sharing our project on 1) accelerating and 2) stabilizing training for large language-vision models 1) Towards accelerating training, we introduce SwitchBack, a linear layer for int8 quantized training which matches bfloat16 within 0.1 for CLIP ViT-Huge arxiv.org/abs/2304.13013
