how to train own vq-f4

[lihao9056]

Great work.But I would like to know how to train own vq-f4.

[Nanicy]

me too

[QiuWYZju]

you can refer first_stage_config of 'ffhq-ldm-vq-4.yaml' and write your own config,such as :

model: base_learning_rate: 4.5e-06 target: taming.models.vqgan.VQModel params: embed_dim: 3 n_embed: 8192 ddconfig: double_z: false z_channels: 3 resolution: 256 in_channels: 3 out_ch: 3 ch: 128 ch_mult: - 1 - 2 - 4 num_res_blocks: 2 attn_resolutions: [] dropout: 0.0 lossconfig: target: taming.modules.losses.vqperceptual.VQLPIPSWithDiscriminator params: disc_conditional: false disc_in_channels: 3 disc_start: 0 disc_weight: 0.8 codebook_weight: 1.0

then rename it as vq-4.yaml. it works when i use the vq-4 yaml file and ckpt in project 'taming-transformers'.