paper11667/clipstyler 📝🖼️🔢 → 🖼️

▶️ 26.0K runs 📅 Dec 2021 ⚙️ Cog dev 🔗 GitHub 📄 Paper ⚖️ License

image-style-transfer image-to-image

About

Image Style Transfer with Text Condition

Example Output

Output

[object Object][object Object][object Object][object Object][object Object]

Performance Metrics

42.09s Prediction Time

68.47s Total Time

All Input Parameters

{
  "text": "sketch with crayon",
  "image": "https://replicate.delivery/mgxm/e4500aa0-f71b-42ff-a540-aadb44c8d1b2/face.jpg",
  "iterations": "100"
}

Input Parameters

text (required) Type: string: text for style transfer
image (required) Type: string: Input image (will be cropped before style transfer)
iterations Type: integerDefault: 100: training iterations

Output Schema

Output

Type: array • Items Type: string • Items Format: uri

Example Execution Logs

/root/.pyenv/versions/3.8.12/lib/python3.8/site-packages/torch/nn/functional.py:3060: UserWarning: Default upsampling behavior when mode=bicubic is changed to align_corners=False since 0.4.0. Please specify align_corners=True if the old behavior is desired. See the documentation of nn.Upsample for details.
  warnings.warn("Default upsampling behavior when mode={} is changed "
/root/.pyenv/versions/3.8.12/lib/python3.8/site-packages/torch/optim/lr_scheduler.py:131: UserWarning: Detected call of `lr_scheduler.step()` before `optimizer.step()`. In PyTorch 1.1.0 and later, you should call them in the opposite order: `optimizer.step()` before `lr_scheduler.step()`.  Failure to do this will result in PyTorch skipping the first value of the learning rate schedule. See more details at https://pytorch.org/docs/stable/optim.html#how-to-adjust-learning-rate
  warnings.warn("Detected call of `lr_scheduler.step()` before `optimizer.step()`. "
/root/.pyenv/versions/3.8.12/lib/python3.8/site-packages/torch/nn/functional.py:3060: UserWarning: Default upsampling behavior when mode=bilinear is changed to align_corners=False since 0.4.0. Please specify align_corners=True if the old behavior is desired. See the documentation of nn.Upsample for details.
  warnings.warn("Default upsampling behavior when mode={} is changed "
/root/.pyenv/versions/3.8.12/lib/python3.8/site-packages/torchvision/transforms/functional_tensor.py:876: UserWarning: Argument fill/fillcolor is not supported for Tensor input. Fill value is zero
  warnings.warn("Argument fill/fillcolor is not supported for Tensor input. Fill value is zero")
After 0 iterations
Total loss: 9802.1005859375
Content loss: 5.383391380310059
patch loss: 0.9482421875
dir loss: 0.91650390625
TV loss: 0.3422693610191345
After 20 iterations
Total loss: 8772.9951171875
Content loss: 3.439800500869751
patch loss: 0.87939453125
dir loss: 0.681640625
TV loss: 0.2753937542438507
After 40 iterations
Total loss: 8309.794921875
Content loss: 2.8816094398498535
patch loss: 0.84228515625
dir loss: 0.5947265625
TV loss: 0.30363643169403076
After 60 iterations
Total loss: 8068.50048828125
Content loss: 2.774502754211426
patch loss: 0.8193359375
dir loss: 0.5517578125
TV loss: 0.3253156840801239
After 80 iterations
Total loss: 7695.32177734375
Content loss: 2.5266687870025635
patch loss: 0.78466796875
dir loss: 0.50390625
TV loss: 0.321286678314209
After 100 iterations
Total loss: 7482.86083984375
Content loss: 2.315229892730713
patch loss: 0.7685546875
dir loss: 0.4384765625
TV loss: 0.32614612579345703

Version Details

Version ID: f2d6b24e6002f25f77ae89c2b0a5987daa6d0bf751b858b94b8416e8542434d1
Version Created: March 16, 2022

Run on Replicate →