Skip to content

Weights & Data

Latest
Latest

Choose a tag to compare

View all tags
@muhd-umer muhd-umer released this 16 Dec 13:25
· 2 commits to main since this release
v0.0.1
a40a8ec
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Verified
Learn about vigilant mode.

CIFAR100 Upscaled

bsrgan_4x_cifar100.zip and swinir_4x_cifar100.zip archives contain ImageFolder style dataset for CIFAR100 upscaled to 128x128 resolution using BSRGAN 4x and SwinIR models, respectively.

Follows the same train/test split as CIFAR100.

BSRGAN 4x

bsrgan_4x.pth contains the BSRGAN 4x model weights. Place the weights in "weights/" folder to use upscale.py script.

EfficientNetV2

tl_effnetv2m_in21k_e50_bsrgan4x.pth and tl_effnetv2m_in21k_e50_swinir4x.pth contain the EfficientNetV2-m model checkpoint trained using transfer learning on BSRGAN 4x CIFAR100 and SwinIR CIFAR100, respectively.

scratch_effnetv2m_mish_cos_e100.ckpt and tl_effnetv2m_in21k_e50.ckpt contains the EfficientNetV2-m model checkpoint trained from scratch and fine-tuned on CIFAR100 using ImageNet 21k weights.

Assets 9
Loading