Update unet/config.json

unet/config.json

@@ -2,17 +2,15 @@
   "_class_name": "UNet2DConditionModel",
   "_diffusers_version": "0.19.0",
   "act_fn": "silu",
-  "addition_embed_type": "text_time",
+  "addition_embed_type": null,
   "addition_embed_type_num_heads": 64,
-  "addition_time_embed_dim": 256,
-  "attention_head_dim": [
-    5,
-    10,
-    20
-  ],
+  "addition_time_embed_dim": null,
+  "attention_head_dim": 8,
+  "attention_type": "default",
   "block_out_channels": [
     320,
     640,
+    1280,
     1280
   ],
   "center_input_sample": false,
@@ -20,14 +18,16 @@
   "class_embeddings_concat": false,
   "conv_in_kernel": 3,
   "conv_out_kernel": 3,
-  "cross_attention_dim": 2048,
+  "cross_attention_dim": 768,
   "cross_attention_norm": null,
   "down_block_types": [
-    "DownBlock2D",
     "CrossAttnDownBlock2D",
-    "CrossAttnDownBlock2D"
+    "CrossAttnDownBlock2D",
+    "CrossAttnDownBlock2D",
+    "DownBlock2D"
   ],
   "downsample_padding": 1,
+  "dropout": 0.0,
   "dual_cross_attention": false,
   "encoder_hid_dim": null,
   "encoder_hid_dim_type": null,
@@ -44,26 +44,24 @@
   "num_class_embeds": null,
   "only_cross_attention": false,
   "out_channels": 4,
-  "projection_class_embeddings_input_dim": 2816,
+  "projection_class_embeddings_input_dim": null,
   "resnet_out_scale_factor": 1.0,
   "resnet_skip_time_act": false,
   "resnet_time_scale_shift": "default",
-  "sample_size": 128,
+  "reverse_transformer_layers_per_block": null,
+  "sample_size": 64,
   "time_cond_proj_dim": null,
   "time_embedding_act_fn": null,
   "time_embedding_dim": null,
   "time_embedding_type": "positional",
   "timestep_post_act": null,
-  "transformer_layers_per_block": [
-    1,
-    2,
-    10
-  ],
+  "transformer_layers_per_block": 1,
   "up_block_types": [
+    "UpBlock2D",
     "CrossAttnUpBlock2D",
     "CrossAttnUpBlock2D",
-    "UpBlock2D"
+    "CrossAttnUpBlock2D"
   ],
-  "upcast_attention": null,
-  "use_linear_projection": true
+  "upcast_attention": false,
+  "use_linear_projection": false
 }