Upload 7 files

configs/inference/bsrnet.yaml

@@ -0,0 +1,8 @@
+target: model.RRDBNet
+params:
+  in_nc: 3
+  out_nc: 3
+  nf: 64
+  nb: 23
+  gc: 32
+  sf: 4

configs/inference/cldm.yaml

@@ -0,0 +1,65 @@
+target: model.ControlLDM
+params:
+  latent_scale_factor: 0.18215
+  unet_cfg:
+    use_checkpoint: True
+    image_size: 32 # unused
+    in_channels: 4
+    out_channels: 4
+    model_channels: 320
+    attention_resolutions: [ 4, 2, 1 ]
+    num_res_blocks: 2
+    channel_mult: [ 1, 2, 4, 4 ]
+    num_head_channels: 64 # need to fix for flash-attn
+    use_spatial_transformer: True
+    use_linear_in_transformer: True
+    transformer_depth: 1
+    context_dim: 1024
+    legacy: False
+  vae_cfg:
+    embed_dim: 4
+    ddconfig:
+      double_z: true
+      z_channels: 4
+      resolution: 256
+      in_channels: 3
+      out_ch: 3
+      ch: 128
+      ch_mult:
+      - 1
+      - 2
+      - 4
+      - 4
+      num_res_blocks: 2
+      attn_resolutions: []
+      dropout: 0.0
+  clip_cfg:
+    embed_dim: 1024
+    vision_cfg:
+      image_size: 224
+      layers: 32
+      width: 1280
+      head_width: 80
+      patch_size: 14
+    text_cfg:
+      context_length: 77
+      vocab_size: 49408
+      width: 1024
+      heads: 16
+      layers: 24
+    layer: "penultimate"
+  controlnet_cfg:
+    use_checkpoint: True
+    image_size: 32 # unused
+    in_channels: 4
+    hint_channels: 4
+    model_channels: 320
+    attention_resolutions: [ 4, 2, 1 ]
+    num_res_blocks: 2
+    channel_mult: [ 1, 2, 4, 4 ]
+    num_head_channels: 64 # need to fix for flash-attn
+    use_spatial_transformer: True
+    use_linear_in_transformer: True
+    transformer_depth: 1
+    context_dim: 1024
+    legacy: False

configs/inference/diffusion.yaml

@@ -0,0 +1,5 @@
+target: model.Diffusion
+params:
+  linear_start: 0.00085
+  linear_end: 0.0120
+  timesteps: 1000

configs/inference/scunet.yaml

@@ -0,0 +1,5 @@
+target: model.SCUNet
+params:
+  in_nc: 3
+  config: [4,4,4,4,4,4,4]
+  dim: 64

configs/inference/swinir.yaml

@@ -0,0 +1,16 @@
+target: model.SwinIR
+params:
+  img_size: 64
+  patch_size: 1
+  in_chans: 3
+  embed_dim: 180
+  depths: [6, 6, 6, 6, 6, 6, 6, 6]
+  num_heads: [6, 6, 6, 6, 6, 6, 6, 6]
+  window_size: 8
+  mlp_ratio: 2
+  sf: 8
+  img_range: 1.0
+  upsampler: "nearest+conv"
+  resi_connection: "1conv"
+  unshuffle: True
+  unshuffle_scale: 8

configs/train/train_stage1.yaml

@@ -0,0 +1,66 @@
+model:
+  swinir:
+    target: model.swinir.SwinIR
+    params:
+      img_size: 64
+      patch_size: 1
+      in_chans: 3
+      embed_dim: 180
+      depths: [6, 6, 6, 6, 6, 6, 6, 6]
+      num_heads: [6, 6, 6, 6, 6, 6, 6, 6]
+      window_size: 8
+      mlp_ratio: 2
+      sf: 8
+      img_range: 1.0
+      upsampler: "nearest+conv"
+      resi_connection: "1conv"
+      unshuffle: True
+      unshuffle_scale: 8
+
+dataset:
+  train:
+    target: dataset.codeformer.CodeformerDataset
+    params:
+      # training file list path
+      file_list: 
+      file_backend_cfg:
+        target: dataset.file_backend.HardDiskBackend
+      out_size: 512
+      crop_type: center
+      blur_kernel_size: 41
+      kernel_list: ['iso', 'aniso']
+      kernel_prob: [0.5, 0.5]
+      blur_sigma: [0.1, 12]
+      downsample_range: [1, 12]
+      noise_range: [0, 15]
+      jpeg_range: [30, 100]
+  val:
+    target: dataset.codeformer.CodeformerDataset
+    params:
+      # validation file list path
+      file_list: 
+      file_backend_cfg:
+        target: dataset.file_backend.HardDiskBackend
+      out_size: 512
+      crop_type: center
+      blur_kernel_size: 41
+      kernel_list: ['iso', 'aniso']
+      kernel_prob: [0.5, 0.5]
+      blur_sigma: [0.1, 12]
+      downsample_range: [1, 12]
+      noise_range: [0, 15]
+      jpeg_range: [30, 100]
+
+train:
+  # experiment directory path
+  exp_dir: 
+  learning_rate: 1e-4
+  # total batch size
+  batch_size: 96
+  num_workers: 
+  train_steps: 150000
+  log_every: 50
+  ckpt_every: 10000
+  image_every: 1000
+  val_every: 1000
+  resume: ~

configs/train/train_stage2.yaml

@@ -0,0 +1,130 @@
+model:
+  cldm:
+    target: model.cldm.ControlLDM
+    params:
+      latent_scale_factor: 0.18215
+      unet_cfg:
+        use_checkpoint: True
+        image_size: 32 # unused
+        in_channels: 4
+        out_channels: 4
+        model_channels: 320
+        attention_resolutions: [ 4, 2, 1 ]
+        num_res_blocks: 2
+        channel_mult: [ 1, 2, 4, 4 ]
+        num_head_channels: 64 # need to fix for flash-attn
+        use_spatial_transformer: True
+        use_linear_in_transformer: True
+        transformer_depth: 1
+        context_dim: 1024
+        legacy: False
+      vae_cfg:
+        embed_dim: 4
+        ddconfig:
+          double_z: true
+          z_channels: 4
+          resolution: 256
+          in_channels: 3
+          out_ch: 3
+          ch: 128
+          ch_mult:
+          - 1
+          - 2
+          - 4
+          - 4
+          num_res_blocks: 2
+          attn_resolutions: []
+          dropout: 0.0
+      clip_cfg:
+        embed_dim: 1024
+        vision_cfg:
+          image_size: 224
+          layers: 32
+          width: 1280
+          head_width: 80
+          patch_size: 14
+        text_cfg:
+          context_length: 77
+          vocab_size: 49408
+          width: 1024
+          heads: 16
+          layers: 24
+        layer: "penultimate"
+      controlnet_cfg:
+        use_checkpoint: True
+        image_size: 32 # unused
+        in_channels: 4
+        hint_channels: 4
+        model_channels: 320
+        attention_resolutions: [ 4, 2, 1 ]
+        num_res_blocks: 2
+        channel_mult: [ 1, 2, 4, 4 ]
+        num_head_channels: 64 # need to fix for flash-attn
+        use_spatial_transformer: True
+        use_linear_in_transformer: True
+        transformer_depth: 1
+        context_dim: 1024
+        legacy: False
+
+  swinir:
+    target: model.swinir.SwinIR
+    params:
+      img_size: 64
+      patch_size: 1
+      in_chans: 3
+      embed_dim: 180
+      depths: [6, 6, 6, 6, 6, 6, 6, 6]
+      num_heads: [6, 6, 6, 6, 6, 6, 6, 6]
+      window_size: 8
+      mlp_ratio: 2
+      sf: 8
+      img_range: 1.0
+      upsampler: "nearest+conv"
+      resi_connection: "1conv"
+      unshuffle: True
+      unshuffle_scale: 8
+
+  diffusion:
+    target: model.gaussian_diffusion.Diffusion
+    params:
+      linear_start: 0.00085
+      linear_end: 0.0120
+      timesteps: 1000
+
+dataset:
+  train:
+    target: dataset.codeformer.CodeformerDataset
+    params:
+      # training file list path
+      file_list: 
+      file_backend_cfg:
+        target: dataset.file_backend.HardDiskBackend
+      out_size: 512
+      crop_type: center
+      blur_kernel_size: 41
+      kernel_list: ['iso', 'aniso']
+      kernel_prob: [0.5, 0.5]
+      blur_sigma: [0.1, 12]
+      downsample_range: [1, 12]
+      noise_range: [0, 15]
+      jpeg_range: [30, 100]
+
+train:
+  # pretrained sd v2.1 path
+  sd_path: 
+  # experiment directory path
+  exp_dir: 
+  # stage 1 swinir path
+  swinir_path: 
+  learning_rate: 1e-4
+  # ImageNet 1k (1.3M images)
+  # batch size = 192, lr = 1e-4, total training steps = 25k
+  # Our filtered laion2b-en (15M images)
+  # batch size = 256, lr = 1e-4 (first 30k), 1e-5 (next 50k), total training steps = 80k
+  batch_size: 256
+  num_workers: 
+  train_steps: 30000
+  log_every: 50
+  ckpt_every: 10000
+  image_every: 1000
+  resume: ~