Pad H3 guide latents to the patch size before patchify

Amp-Thread-ID: https://ampcode.com/threads/T-019ff272-8eec-75bb-9297-07d66ae9c739
Co-authored-by: Amp <amp@ampcode.com>

comfy/ldm/minimax/model.py

@@ -485,6 +485,7 @@ class MiniMaxH3Model(nn.Module):
         seed = int(payload.get("seed", 0))
         # every condition intentionally restarts the same RNG stream
         for z in payload.get("cond_video_latents", []):
+            z = comfy.ldm.common_dit.pad_to_patch_size(z, self.patch_size)
             r = patchify_video(z.to(torch.float32), self.patch_size)
             if aug < 1.0:
                 gen = torch.Generator("cpu").manual_seed(seed)