update train_unconditional for latent diffusion

scripts/train_vae.py

@@ -3,6 +3,7 @@
 # convert_original_stable_diffusion_to_diffusers.py
 
 # TODO
+# add latent resolution as parameter
 # grayscale
 # update README
 
@@ -151,7 +152,7 @@ if __name__ == "__main__":
         trainer_opt,
         resume_from_checkpoint=args.resume_from_checkpoint,
         callbacks=[
-            ImageLogger(every=10),
+            ImageLogger(),
             HFModelCheckpoint(ldm_config=config,
                               hf_checkpoint=args.hf_checkpoint_dir,
                               dirpath=args.ldm_checkpoint_dir,