| { | |
| "bias_init": "zeros", | |
| "compute_dropout": 0, | |
| "compute_heads": 16, | |
| "compute_stochastic_depth": 0, | |
| "concat_cond_token_to_latents": false, | |
| "data_dim": 256, | |
| "data_mlp_multiplier": 4, | |
| "data_positional_embedding_type": "learned", | |
| "data_size": 32, | |
| "label_dim": 2048, | |
| "latent_mlp_multiplier": 4, | |
| "latents_dim": 1024, | |
| "modulate_conditioning": false, | |
| "noise_embedding_type": "positional", | |
| "num_blocks": 4, | |
| "num_cond_tokens": 77, | |
| "num_input_channels": 4, | |
| "num_latents": 256, | |
| "num_processing_layers": 4, | |
| "num_text_registers": 16, | |
| "path_size": 2, | |
| "read_write_heads": 16, | |
| "rw_dropout": 0.0, | |
| "rw_stochastic_depth": 0, | |
| "time_scaling": 1000.0, | |
| "use_16_bits_layer_norm": false, | |
| "use_biases": true, | |
| "use_cond_rin_block": true, | |
| "use_cond_token": true, | |
| "weight_init": "xavier_uniform" | |
| } |