Commit History

  1. Add model card

    Verified signature
    2853c71

  1. Add wandb logging (Life is more peacefull)

    7e6af10

  1. Model not moving forward

    bf96a55
  2. Add update logic

    2e5aaca
  3. Add scheduler

    04897d8
  4. Sloving conflicts

    88f49f7
  5. More stuff

    fca91c0
  6. Update disc_loss update logic

    eec35f5
  7. Made changes in the training script

    2c4ef68
  8. More samples

    90d71da
  9. Update

    f4c7526
  10. Re training will finish this time for sure

    8082844

  1. More samples

    680c645
  2. Halfway through

    de7e31f
  3. Remove all samples

    e183f2c

  1. Checkpoint for 2nd epoch

    1c6f5f9
  2. Checkpointnt for 1st epoch

    af8cdd6
  3. Change checkpoint dir

    1a70dc1
  4. Modify training code

    bba6806
  5. Remove logging statements

    2473850
  6. Move to gpu and add checkpoint images to attributes

    f5927bd
  7. Print statements for logging

    43b6c80
  8. Add num_heads variables

    Yash Nagraj
    92078be
  9. Make changes

    Yash Nagraj
    f0ff580
  10. Resolve run time errors

    8d85e1f
  11. Make changes to get image

    Yash Nagraj
    789dd08
  12. Get the buffers in order to open as image

    Yash Nagraj
    199f027
  13. Change any output errors

    788a6e1
  14. Add im_path from the config

    Yash Nagraj
    1013ca2
  15. Change the config file accordingly

    Yash Nagraj
    432f436

  1. Add loss printing and saving logic

    Yash Nagraj
    936be8c
  2. Optimize Discriminator in the training loop

    Yash Nagraj
    e57e318
  3. Change dataloader loading from directory

    Yash Nagraj
    d67fec6

  1. Add dataloader for training

    Yash Nagraj
    4c5a8ab
  2. Add the dataset

    Yash Nagraj
    2b3a313
  3. Move the pretrained vgg16 model to another directory

    Yash Nagraj
    f6182a3
  4. Add forward function to the Diffusion model UNet

    Yash Nagraj
    dff68d4
  5. Build the Unet from the config

    Yash Nagraj
    6966d99
  6. change variable names

    Yash Nagraj
    2e8c272
  7. Get all parameters from model_config

    Yash Nagraj
    97679c9
  8. Add discriminator for adverserial loss for VAE training

    Yash Nagraj
    ec937a4

  1. Add lpips for VAE training

    Yash Nagraj
    b311ae5
  2. Clean up

    Yash Nagraj
    2a1c7f2
  3. Add AutoEncoder (VQVAE)

    Yash Nagraj
    7377e9c

  1. Add UpBlock Unet to concat downblock's output

    Yash Nagraj
    70a401a
  2. Add all the layers in the UpBlock

    Yash Nagraj
    6801d5a

  1. Add forward function to the midblock

    Yash Nagraj
    de5e356
  2. Add residual convolution to the mid block

    Yash Nagraj
    38d054a

  1. Add MidBlocks and change cross_attn in down blocks

    Yash Nagraj
    c2857b5
  2. Add cross attn if needed for conditional latent diffusion

    Yash Nagraj
    95b2cf2