update diffusers weights

image_encoder/config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "StableCascade-prior/image_encoder",
+  "_name_or_path": "openai/clip-vit-large-patch14",
   "architectures": [
     "CLIPVisionModelWithProjection"
   ],
@@ -18,6 +18,6 @@
   "num_hidden_layers": 24,
   "patch_size": 14,
   "projection_dim": 768,
-  "torch_dtype": "bfloat16",
-  "transformers_version": "4.38.0.dev0"
+  "torch_dtype": "float32",
+  "transformers_version": "4.38.2"
 }

image_encoder/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e4b33d864f89a793357a768cb07d0dc18d6a14e6664f4110a0d535ca9ba78da8
-size 607980488
+oid sha256:77b33d2a3a643650857672e880ccf73adbaf114fbbadec36d142ee9d48af7e20
+size 1215912728

model_index.json

@@ -1,7 +1,6 @@
 {
   "_class_name": "StableCascadePriorPipeline",
-  "_diffusers_version": "0.26.0.dev0",
-  "_name_or_path": "StableCascade-prior/",
+  "_diffusers_version": "0.27.0.dev0",
   "feature_extractor": [
     "transformers",
     "CLIPImageProcessor"
@@ -11,8 +10,8 @@
     "CLIPVisionModelWithProjection"
   ],
   "prior": [
-    "stable_cascade",
-    "StableCascadeUnet"
+    "diffusers",
+    "StableCascadeUNet"
   ],
   "resolution_multiple": 42.67,
   "scheduler": [

prior/config.json

@@ -1,61 +1,64 @@
 {
-  "_class_name": "StableCascadeUnet",
-  "_diffusers_version": "0.26.0.dev0",
-  "_name_or_path": "StableCascade-prior/prior",
-  "block_repeat": [
-    [
-      1,
-      1
-    ],
-    [
-      1,
-      1
-    ]
+  "_class_name": "StableCascadeUNet",
+  "_diffusers_version": "0.27.0.dev0",
+  "block_out_channels": [
+    2048,
+    2048
   ],
-  "blocks": [
+  "block_types_per_layer": [
     [
-      8,
-      24
+      "SDCascadeResBlock",
+      "SDCascadeTimestepBlock",
+      "SDCascadeAttnBlock"
     ],
     [
-      24,
-      8
+      "SDCascadeResBlock",
+      "SDCascadeTimestepBlock",
+      "SDCascadeAttnBlock"
     ]
   ],
-  "c_clip_img": 768,
-  "c_clip_seq": 4,
-  "c_clip_text": 1280,
-  "c_clip_text_pooled": 1280,
-  "c_cond": 2048,
-  "c_effnet": null,
-  "c_hidden": [
-    2048,
-    2048
+  "clip_image_in_channels": 768,
+  "clip_seq": 4,
+  "clip_text_in_channels": 1280,
+  "clip_text_pooled_in_channels": 1280,
+  "conditioning_dim": 2048,
+  "down_blocks_repeat_mappers": [
+    1,
+    1
+  ],
+  "down_num_layers_per_block": [
+    8,
+    24
   ],
-  "c_in": 16,
-  "c_out": 16,
-  "c_pixels": null,
-  "c_r": 64,
   "dropout": [
     0.1,
     0.1
   ],
+  "effnet_in_channels": null,
+  "in_channels": 16,
   "kernel_size": 3,
-  "level_config": [
-    "CTA",
-    "CTA"
-  ],
-  "nhead": [
+  "num_attention_heads": [
     32,
     32
   ],
+  "out_channels": 16,
   "patch_size": 1,
+  "pixel_mapper_in_channels": null,
   "self_attn": true,
   "switch_level": [
     false
   ],
-  "t_conds": [
+  "timestep_conditioning_type": [
     "sca",
     "crp"
+  ],
+  "timestep_ratio_embedding_dim": 64,
+  "up_blocks_repeat_mappers": [
+    1,
+    1
+  ],
+  "up_num_layers_per_block": [
+    24,
+    8
   ]
 }

prior/diffusion_pytorch_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:44a4cd9540f327f2fb4ac09179e4e87912a01cdb1b3b86c79f0f853976fb4c98
-size 7178377816
+oid sha256:0a2c7aa62c503780b85f74fd513b1b99c12ea4f83422bdbad5ac264aa68efb4b
+size 14356584672

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-prior/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": "4.38.2",
   "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/tokenizer.json

@@ -1,21 +1,7 @@
 {
   "version": "1.0",
-  "truncation": {
-    "direction": "Right",
-    "max_length": 77,
-    "strategy": "LongestFirst",
-    "stride": 0
-  },
-  "padding": {
-    "strategy": {
-      "Fixed": 77
-    },
-    "direction": "Right",
-    "pad_to_multiple_of": null,
-    "pad_id": 49407,
-    "pad_type_id": 0,
-    "pad_token": "<|endoftext|>"
-  },
+  "truncation": null,
+  "padding": null,
   "added_tokens": [
     {
       "id": 49406,