RunDiffusion/RunDiffusion-XL-Beta

1{2  "_class_name": "UNet2DConditionModel",3  "_diffusers_version": "0.20.0.dev0",4  "act_fn": "silu",5  "addition_embed_type": "text_time",6  "addition_embed_type_num_heads": 64,7  "addition_time_embed_dim": 256,8  "attention_head_dim": [9    5,10    10,11    2012  ],13  "block_out_channels": [14    320,15    640,16    128017  ],18  "center_input_sample": false,19  "class_embed_type": null,20  "class_embeddings_concat": false,21  "conv_in_kernel": 3,22  "conv_out_kernel": 3,23  "cross_attention_dim": 2048,24  "cross_attention_norm": null,25  "down_block_types": [26    "DownBlock2D",27    "CrossAttnDownBlock2D",28    "CrossAttnDownBlock2D"29  ],30  "downsample_padding": 1,31  "dual_cross_attention": false,32  "encoder_hid_dim": null,33  "encoder_hid_dim_type": null,34  "flip_sin_to_cos": true,35  "freq_shift": 0,36  "in_channels": 4,37  "layers_per_block": 2,38  "mid_block_only_cross_attention": null,39  "mid_block_scale_factor": 1,40  "mid_block_type": "UNetMidBlock2DCrossAttn",41  "norm_eps": 1e-05,42  "norm_num_groups": 32,43  "num_attention_heads": null,44  "num_class_embeds": null,45  "only_cross_attention": false,46  "out_channels": 4,47  "projection_class_embeddings_input_dim": 2816,48  "resnet_out_scale_factor": 1.0,49  "resnet_skip_time_act": false,50  "resnet_time_scale_shift": "default",51  "sample_size": 128,52  "time_cond_proj_dim": null,53  "time_embedding_act_fn": null,54  "time_embedding_dim": null,55  "time_embedding_type": "positional",56  "timestep_post_act": null,57  "transformer_layers_per_block": [58    1,59    2,60    1061  ],62  "up_block_types": [63    "CrossAttnUpBlock2D",64    "CrossAttnUpBlock2D",65    "UpBlock2D"66  ],67  "upcast_attention": null,68  "use_linear_projection": true69}70