From a02e6d93855777f1c04fd778b05b0f76792b07f2 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Gabriel=20Ilharco=20Magalh=C3=A3es?= Date: Tue, 3 Aug 2021 22:28:52 -0700 Subject: [PATCH] Adding ResNet-101 checkpoint --- README.md | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/README.md b/README.md index 92bed39c2..dc13fcee8 100644 --- a/README.md +++ b/README.md @@ -111,7 +111,10 @@ python src/training/main.py \ ## Trained models -You can find our ResNet-50 trained on YFCC-15M [here](https://drive.google.com/file/d/1bbNMrWAq3NxCAteHmbrYgBKpGAA9j9pn/view?usp=sharing). +Below are checkpoints of models trained on YFCC-15M, along with their zero-shot top-1 accuracies on ImageNet and ImageNetV2. These models were trained using 8 GPUs and the same hyperparameters described in the "Sample running code" section, with the exception of `lr=5e-4` and `epochs=32`. + +* [ResNet-50](https://drive.google.com/file/d/1bbNMrWAq3NxCAteHmbrYgBKpGAA9j9pn/view?usp=sharing) (32.7% / 27.9%) +* [ResNet-101](https://drive.google.com/file/d/1vOorR3pKkuuA_Gv7OgqMtaETNjTmy_3m/view?usp=sharing) (34.8% / 30.0%) ## Scaling trends