Add Diffusers weights (#5)

- Add Diffusers weights (2e36a6dc7bcde2158678496699630a2e6dd5d132)


Co-authored-by: Alex Poulos <[email protected]>

feature_extractor/preprocessor_config.json

@@ -1,8 +1,12 @@
 {
-  "crop_size": 224,
+  "crop_size": {
+    "height": 224,
+    "width": 224
+  },
   "do_center_crop": true,
   "do_convert_rgb": true,
   "do_normalize": true,
+  "do_rescale": true,
   "do_resize": true,
   "feature_extractor_type": "CLIPFeatureExtractor",
   "image_mean": [
@@ -10,11 +14,15 @@
     0.4578275,
     0.40821073
   ],
+  "image_processor_type": "CLIPImageProcessor",
   "image_std": [
     0.26862954,
     0.26130258,
     0.27577711
   ],
   "resample": 3,
-  "size": 224
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "shortest_edge": 224
+  }
 }

model_index.json

@@ -1,10 +1,11 @@
 {
   "_class_name": "StableDiffusionPipeline",
-  "_diffusers_version": "0.8.0.dev0",
+  "_diffusers_version": "0.10.0.dev0",
   "feature_extractor": [
     "transformers",
-    "CLIPFeatureExtractor"
+    "CLIPImageProcessor"
   ],
+  "requires_safety_checker": true,
   "safety_checker": [
     "stable_diffusion",
     "StableDiffusionSafetyChecker"

safety_checker/config.json

@@ -1,5 +1,5 @@
 {
-  "_commit_hash": "4bb648a606ef040e7685bde262611766a5fdd67b",
+  "_commit_hash": "cb41f3a270d63d454d385fc2e4f571c487c253c5",
   "_name_or_path": "CompVis/stable-diffusion-safety-checker",
   "architectures": [
     "StableDiffusionSafetyChecker"
@@ -62,6 +62,7 @@
     "pad_token_id": 1,
     "prefix": null,
     "problem_type": null,
+    "projection_dim": 512,
     "pruned_heads": {},
     "remove_invalid_values": false,
     "repetition_penalty": 1.0,
@@ -79,7 +80,7 @@
     "top_p": 1.0,
     "torch_dtype": null,
     "torchscript": false,
-    "transformers_version": "4.24.0",
+    "transformers_version": "4.26.0.dev0",
     "typical_p": 1.0,
     "use_bfloat16": false,
     "vocab_size": 49408
@@ -148,6 +149,7 @@
     "patch_size": 14,
     "prefix": null,
     "problem_type": null,
+    "projection_dim": 512,
     "pruned_heads": {},
     "remove_invalid_values": false,
     "repetition_penalty": 1.0,
@@ -165,7 +167,7 @@
     "top_p": 1.0,
     "torch_dtype": null,
     "torchscript": false,
-    "transformers_version": "4.24.0",
+    "transformers_version": "4.26.0.dev0",
     "typical_p": 1.0,
     "use_bfloat16": false
   },

scheduler/scheduler_config.json

@@ -1,10 +1,12 @@
 {
   "_class_name": "PNDMScheduler",
-  "_diffusers_version": "0.8.0.dev0",
+  "_diffusers_version": "0.10.0.dev0",
   "beta_end": 0.012,
   "beta_schedule": "scaled_linear",
   "beta_start": 0.00085,
+  "clip_sample": false,
   "num_train_timesteps": 1000,
+  "prediction_type": "epsilon",
   "set_alpha_to_one": false,
   "skip_prk_steps": true,
   "steps_offset": 1,

text_encoder/config.json

@@ -20,6 +20,6 @@
   "pad_token_id": 1,
   "projection_dim": 768,
   "torch_dtype": "float32",
-  "transformers_version": "4.24.0",
+  "transformers_version": "4.26.0.dev0",
   "vocab_size": 49408
 }

unet/config.json

@@ -1,6 +1,6 @@
 {
   "_class_name": "UNet2DConditionModel",
-  "_diffusers_version": "0.8.0.dev0",
+  "_diffusers_version": "0.10.0.dev0",
   "act_fn": "silu",
   "attention_head_dim": 8,
   "block_out_channels": [
@@ -18,6 +18,7 @@
     "DownBlock2D"
   ],
   "downsample_padding": 1,
+  "dual_cross_attention": false,
   "flip_sin_to_cos": true,
   "freq_shift": 0,
   "in_channels": 4,
@@ -25,6 +26,8 @@
   "mid_block_scale_factor": 1,
   "norm_eps": 1e-05,
   "norm_num_groups": 32,
+  "num_class_embeds": null,
+  "only_cross_attention": false,
   "out_channels": 4,
   "sample_size": 64,
   "up_block_types": [
@@ -32,5 +35,6 @@
     "CrossAttnUpBlock2D",
     "CrossAttnUpBlock2D",
     "CrossAttnUpBlock2D"
-  ]
+  ],
+  "use_linear_projection": false
 }

vae/config.json

@@ -1,6 +1,6 @@
 {
   "_class_name": "AutoencoderKL",
-  "_diffusers_version": "0.8.0.dev0",
+  "_diffusers_version": "0.10.0.dev0",
   "act_fn": "silu",
   "block_out_channels": [
     128,
@@ -19,7 +19,7 @@
   "layers_per_block": 2,
   "norm_num_groups": 32,
   "out_channels": 3,
-  "sample_size": 256,
+  "sample_size": 512,
   "up_block_types": [
     "UpDecoderBlock2D",
     "UpDecoderBlock2D",