Upload 12 files

models/CLIP/asian_composite_clip.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:da5a94aaafb8dffcad3ebd0d698933416f82a186a0140425f326c13ca1c07cb0
+size 1144165570

models/configs/celeba_config.yaml

@@ -0,0 +1,33 @@
+encoder:
+    in_channels: 3
+    resolution: 128
+    channels: 128
+    channels_multipliers: [1, 2, 4]
+    z_channels: 3
+    num_res_blocks: 2
+    dropout: 0.0
+    
+decoder:
+    in_channels: 3
+    out_channels: 3
+    resolution: 128
+    channels: 128
+    channels_multipliers: [1, 2, 4]
+    z_channels: 6
+    num_res_blocks: 2
+    dropout: 0.0
+
+vae:
+    embed_dim: 3
+    kld_weight: 1e-6
+
+sampler:
+    beta: 'linear'
+    T: 1000
+    ddim_T: 100
+    ddim_eta: 1
+
+unet:
+    dim: 224
+    dim_mults: [1, 2, 3, 4]
+    channels: 3

models/configs/cifar10_config.yaml

@@ -0,0 +1,33 @@
+encoder:
+    in_channels: 3
+    resolution: 32
+    channels: 128
+    channels_multipliers: [1, 2,]
+    z_channels: 3
+    num_res_blocks: 2
+    dropout: 0.0
+    
+decoder:
+    in_channels: 3
+    out_channels: 3
+    resolution: 32
+    channels: 128
+    channels_multipliers: [1, 2,]
+    z_channels: 6
+    num_res_blocks: 2
+    dropout: 0.0
+
+vae:
+    embed_dim: 3
+    kld_weight: 1e-6
+
+sampler:
+    beta: 'linear'
+    T: 1000
+    ddim_T: 100
+    ddim_eta: 1
+
+unet:
+    dim: 224
+    dim_mults: [1, 2, 3, 4]
+    channels: 3

models/configs/composite_clip_config.yaml

@@ -0,0 +1,15 @@
+text_encoder:
+    number_of_features: 512
+    number_of_heads: 8
+    number_of_transformer_layers: 6
+    context_length: 77
+    embed_dim: 512
+    
+image_encoder:
+    in_channels: 3
+    resolution: 256
+    patch_size: 32
+    number_of_features: 768
+    number_of_heads: 12
+    number_of_transformer_layers: 6
+    embed_dim: 512

models/configs/composite_config.yaml

@@ -0,0 +1,35 @@
+encoder:
+    in_channels: 3
+    resolution: 256
+    channels: 128
+    channels_multipliers: [1, 2, 4]
+    z_channels: 3
+    num_res_blocks: 2
+    dropout: 0.0
+    
+decoder:
+    in_channels: 3
+    out_channels: 3
+    resolution: 256
+    channels: 128
+    channels_multipliers: [1, 2, 4]
+    z_channels: 6
+    num_res_blocks: 2
+    dropout: 0.0
+
+vae:
+    embed_dim: 3
+    kld_weight: 1e-6
+
+sampler:
+    beta: 'linear'
+    T: 1000
+    ddim_T: 100
+    ddim_eta: 1
+
+unet:
+    dim: 256
+    dim_mults: [1, 2, 3, 4]
+    channels: 3
+    latent_dim: 3
+    text_dim: 512

models/diffusion_models/asian_composite_dm_with_clip.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c9aea96f18ee416827c2d64275c724edc919666b90a49ca91dd92bbe94562f27
+size 4838220794

models/diffusion_models/asian_composite_dm_with_koclip.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:39756443a8926423dfe945fe3222726b0f9de280a0d578c729f9ee4c0109881f
+size 7667082930

models/diffusion_models/celeba_dm.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eb7aa6e68885d25be62bb080cf15ce93fd651a84d5ecee9a81e56e341d2dc3ff
+size 3133237417

models/diffusion_models/cifar_dm.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:97a23d78979410c490de33b8bf5d91a990fa7baec1ba5cda9cd83a2102b2a9a8
+size 2858641545

models/vae/asian_composite_vae.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3e6e9dcedab8380b6d2ab811d7d584fa52b7b6e4a25e5866af1e1544eb7817d6
+size 721608906

models/vae/celeba_vae.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ec6348e3d5a8f5db2b541170ecfa7cbfc01cd8dcb15d51ba5c3d2ccb49d4f9a7
+size 721603114

models/vae/cifar_vae.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:36de480b5b4b3c80d6f4a0c2a01a9fb51e2b19a7da91d0f13f732fd1d0f8c7f3
+size 172423562