heheyas
init
cfb7702
model:
target: sgm.models.diffusion.DiffusionEngine
params:
scale_factor: 0.18215
disable_first_stage_autocast: True
denoiser_config:
target: sgm.modules.diffusionmodules.denoiser.DiscreteDenoiser
params:
num_idx: 1000
scaling_config:
target: sgm.modules.diffusionmodules.denoiser_scaling.VScaling
discretization_config:
target: sgm.modules.diffusionmodules.discretizer.LegacyDDPMDiscretization
network_config:
target: sgm.modules.diffusionmodules.openaimodel.UNetModel
params:
use_checkpoint: True
in_channels: 4
out_channels: 4
model_channels: 320
attention_resolutions: [4, 2, 1]
num_res_blocks: 2
channel_mult: [1, 2, 4, 4]
num_head_channels: 64
use_linear_in_transformer: True
transformer_depth: 1
context_dim: 1024
conditioner_config:
target: sgm.modules.GeneralConditioner
params:
emb_models:
- is_trainable: False
input_key: txt
target: sgm.modules.encoders.modules.FrozenOpenCLIPEmbedder
params:
freeze: true
layer: penultimate
first_stage_config:
target: sgm.models.autoencoder.AutoencoderKL
params:
embed_dim: 4
monitor: val/rec_loss
ddconfig:
double_z: true
z_channels: 4
resolution: 256
in_channels: 3
out_ch: 3
ch: 128
ch_mult: [1, 2, 4, 4]
num_res_blocks: 2
attn_resolutions: []
dropout: 0.0
lossconfig:
target: torch.nn.Identity