update diffusers weights

decoder/config.json

@@ -1,74 +1,83 @@
 {
-  "_class_name": "StableCascadeUnet",
-  "_diffusers_version": "0.26.0.dev0",
-  "_name_or_path": "StableCascade/decoder",
-  "block_repeat": [
+  "_class_name": "StableCascadeUNet",
+  "_diffusers_version": "0.27.0.dev0",
+  "block_out_channels": [
+    320,
+    640,
+    1280,
+    1280
+  ],
+  "block_types_per_layer": [
     [
-      1,
-      1,
-      1,
-      1
+      "SDCascadeResBlock",
+      "SDCascadeTimestepBlock"
     ],
     [
-      3,
-      3,
-      2,
-      2
-    ]
-  ],
-  "blocks": [
+      "SDCascadeResBlock",
+      "SDCascadeTimestepBlock"
+    ],
     [
-      2,
-      6,
-      28,
-      6
+      "SDCascadeResBlock",
+      "SDCascadeTimestepBlock",
+      "SDCascadeAttnBlock"
     ],
     [
-      6,
-      28,
-      6,
-      2
+      "SDCascadeResBlock",
+      "SDCascadeTimestepBlock",
+      "SDCascadeAttnBlock"
     ]
   ],
-  "c_clip_img": null,
-  "c_clip_seq": 4,
-  "c_clip_text": null,
-  "c_clip_text_pooled": 1280,
-  "c_cond": 1280,
-  "c_effnet": 16,
-  "c_hidden": [
-    320,
-    640,
-    1280,
-    1280
+  "clip_image_in_channels": null,
+  "clip_seq": 4,
+  "clip_text_in_channels": null,
+  "clip_text_pooled_in_channels": 1280,
+  "conditioning_dim": 1280,
+  "down_blocks_repeat_mappers": [
+    1,
+    1,
+    1,
+    1
+  ],
+  "down_num_layers_per_block": [
+    2,
+    6,
+    28,
+    6
   ],
-  "c_in": 4,
-  "c_out": 4,
-  "c_pixels": 3,
-  "c_r": 64,
   "dropout": [
     0,
     0,
     0.1,
     0.1
   ],
+  "effnet_in_channels": 16,
+  "in_channels": 4,
   "kernel_size": 3,
-  "level_config": [
-    "CT",
-    "CT",
-    "CTA",
-    "CTA"
-  ],
-  "nhead": [
-    -1,
-    -1,
+  "num_attention_heads": [
+    0,
+    0,
     20,
     20
   ],
+  "out_channels": 4,
   "patch_size": 2,
+  "pixel_mapper_in_channels": 3,
   "self_attn": true,
   "switch_level": null,
-  "t_conds": [
+  "timestep_conditioning_type": [
     "sca"
+  ],
+  "timestep_ratio_embedding_dim": 64,
+  "up_blocks_repeat_mappers": [
+    3,
+    3,
+    2,
+    2
+  ],
+  "up_num_layers_per_block": [
+    6,
+    28,
+    6,
+    2
   ]
 }

decoder/diffusion_pytorch_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1f9575dfa6c2535ad65733d6257d17a7b1e1b54b7eafb251ce9556595f3bc0c9
-size 3126071088
+oid sha256:91e8b7d97350689a234bee74e53325e54191ac4146f599cff9551227566a46cb
+size 6251952232

model_index.json

@@ -1,10 +1,9 @@
 {
   "_class_name": "StableCascadeDecoderPipeline",
-  "_diffusers_version": "0.26.0.dev0",
-  "_name_or_path": "StableCascade/",
+  "_diffusers_version": "0.27.0.dev0",
   "decoder": [
-    "stable_cascade",
-    "StableCascadeUnet"
+    "diffusers",
+    "StableCascadeUNet"
   ],
   "latent_dim_scale": 10.67,
   "scheduler": [

scheduler/scheduler_config.json

@@ -1,6 +1,6 @@
 {
   "_class_name": "DDPMWuerstchenScheduler",
-  "_diffusers_version": "0.26.0.dev0",
+  "_diffusers_version": "0.27.0.dev0",
   "s": 0.008,
   "scaler": 1.0
 }

text_encoder/config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "StableCascade/text_encoder",
+  "_name_or_path": "laion/CLIP-ViT-bigG-14-laion2B-39B-b160k",
   "architectures": [
     "CLIPTextModelWithProjection"
   ],
@@ -19,7 +19,7 @@
   "num_hidden_layers": 32,
   "pad_token_id": 1,
   "projection_dim": 1280,
-  "torch_dtype": "bfloat16",
-  "transformers_version": "4.38.0.dev0",
+  "torch_dtype": "float32",
+  "transformers_version": "4.35.2",
   "vocab_size": 49408
 }

text_encoder/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:260e0127aca3c89db813637ae659ebb822cb07af71fedc16cbd980e9518dfdcd
-size 1389382688
+oid sha256:fa5b2e6f4c2efc2d82e4b8312faec1a5540eabfc6415126c9a05c8436a530ef4
+size 2778702264

tokenizer/special_tokens_map.json

@@ -9,21 +9,15 @@
   "eos_token": {
     "content": "<|endoftext|>",
     "lstrip": false,
-    "normalized": false,
-    "rstrip": false,
-    "single_word": false
-  },
-  "pad_token": {
-    "content": "<|endoftext|>",
-    "lstrip": false,
-    "normalized": false,
+    "normalized": true,
     "rstrip": false,
     "single_word": false
   },
+  "pad_token": "<|endoftext|>",
   "unk_token": {
     "content": "<|endoftext|>",
     "lstrip": false,
-    "normalized": false,
+    "normalized": true,
     "rstrip": false,
     "single_word": false
   }

tokenizer/tokenizer.json

@@ -1,14 +1,7 @@
 {
   "version": "1.0",
   "truncation": null,
-  "padding": {
-    "strategy": "BatchLongest",
-    "direction": "Right",
-    "pad_to_multiple_of": null,
-    "pad_id": 49407,
-    "pad_type_id": 0,
-    "pad_token": "<|endoftext|>"
-  },
+  "padding": null,
   "added_tokens": [
     {
       "id": 49406,
@@ -25,7 +18,7 @@
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
-      "normalized": false,
+      "normalized": true,
       "special": true
     }
   ],

tokenizer/tokenizer_config.json

@@ -12,7 +12,7 @@
     "49407": {
       "content": "<|endoftext|>",
       "lstrip": false,
-      "normalized": false,
+      "normalized": true,
       "rstrip": false,
       "single_word": false,
       "special": true

vqgan/config.json

@@ -1,7 +1,7 @@
 {
   "_class_name": "PaellaVQModel",
-  "_diffusers_version": "0.26.0.dev0",
-  "_name_or_path": "StableCascade/vqgan",
+  "_diffusers_version": "0.27.0.dev0",
+  "_name_or_path": "warp-ai/wuerstchen",
   "bottleneck_blocks": 12,
   "embed_dim": 384,
   "in_channels": 3,

vqgan/diffusion_pytorch_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3ac32fab5177329dac907b2480c8c00aeefc712dfd92c2d52263a9c64b426b26
-size 36825828
+oid sha256:052db8852c0d8b117e6d2a59ae3e0c7d7aaae3d00f247e392ef8e9837e11d6c4
+size 73639568