Upload folder using huggingface_hub

README.md

@@ -58,6 +58,10 @@ image.save("output.png")
 ## Conversion Details
 
 The model was converted from ComfyUI format to diffusers format with the following key transformations:
-- Split combined `attention.qkv` weights into separate `to_q`, `to_k`, `to_v` 
+- Removed `model.diffusion_model.` prefix from all keys
+- Renamed `x_embedder` to `all_x_embedder.2-1`
+- Renamed `final_layer` to `all_final_layer.2-1`
+- Split `attention.qkv` into `attention.to_q`, `attention.to_k`, `attention.to_v`
 - Renamed `attention.out` to `attention.to_out.0`
-- Renamed normalization layers to match diffusers conventions
+- Renamed `attention.q_norm` to `attention.norm_q`
+- Renamed `attention.k_norm` to `attention.norm_k`

config.json

@@ -1,16 +1,31 @@
 {
   "_class_name": "ZImageTransformer2DModel",
-  "_diffusers_version": "0.32.0",
-  "attention_head_dim": 128,
-  "cap_embed_dim": 4096,
+  "_diffusers_version": "0.36.0.dev0",
+  "all_f_patch_size": [
+    1
+  ],
+  "all_patch_size": [
+    2
+  ],
+  "axes_dims": [
+    32,
+    48,
+    48
+  ],
+  "axes_lens": [
+    1536,
+    512,
+    512
+  ],
+  "cap_feat_dim": 2560,
+  "dim": 3840,
   "in_channels": 16,
-  "joint_attention_dim": 4096,
-  "mlp_ratio": 4.0,
-  "num_attention_heads": 32,
-  "num_layers": 30,
-  "out_channels": 16,
-  "patch_size": 2,
-  "pooled_projection_dim": 4096,
-  "qk_norm": "rms_norm",
-  "text_embed_dim": 4096
+  "n_heads": 30,
+  "n_kv_heads": 30,
+  "n_layers": 30,
+  "n_refiner_layers": 2,
+  "norm_eps": 1e-05,
+  "qk_norm": true,
+  "rope_theta": 256.0,
+  "t_scale": 1000.0
 }

diffusion_pytorch_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:afc6f40ec35da524055e78d09b8516d53222ed499dab0a685f4eb5f9df1f2cbb
-size 12309874192
+oid sha256:2002fbb39c9790b408efc6fd3f92f5ed25e68b4e94d347231de992f78ee68968
+size 12309881888