transformer/config.json

{
  "_class_name": "Transformer2DModel",
  "_diffusers_version": "0.32.1",
  "attention_head_dim": 128,
  "axes_dims_rope": [
    16,
    56,
    56
  ],
  "codebook_size": 8192,
  "connector_type": "linear",
  "downsample": true,
  "guidance_embeds": false,
  "in_channels": 64,
  "joint_attention_dim": 2304,
  "num_attention_heads": 8,
  "num_layers": 14,
  "num_single_layers": 28,
  "patch_size": 1,
  "pooled_projection_dim": 1024,
  "text_encoder_architecture": "CLIP_Gemma2",
  "upsample": true,
  "vocab_size": 8256
}