| { | |
| "latent_channels": 96, | |
| "resolution": 256, | |
| "encoder_widths": [ | |
| 120, | |
| 240, | |
| 376, | |
| 496 | |
| ], | |
| "encoder_attention": 2, | |
| "encoder_attention_f16": 2, | |
| "decoder_width": 768, | |
| "decoder_depth": 6, | |
| "decoder_heads": 12, | |
| "hires_width": 768, | |
| "hires_depth": 6, | |
| "hires_heads": 12, | |
| "refiner_width": 48, | |
| "refiner2_width": 96, | |
| "layer_scale": 0.0001, | |
| "training_step": 400000 | |
| } | |