diff --git a/modules/processing_diffusers.py b/modules/processing_diffusers.py
index 335122b45..6fa8cec24 100644
--- a/modules/processing_diffusers.py
+++ b/modules/processing_diffusers.py
@@ -86,7 +86,6 @@ def process_diffusers(p: processing.StableDiffusionProcessing):
         denoising_start=0 if use_refiner_start else p.refiner_start if use_denoise_start else None,
         denoising_end=p.refiner_start if use_refiner_start else 1 if use_denoise_start else None,
         output_type='latent' if hasattr(shared.sd_model, 'vae') else 'np',
-        # output_type='pil',
         clip_skip=p.clip_skip,
         desc='Base',
     )
diff --git a/modules/processing_vae.py b/modules/processing_vae.py
index e8bea9a8c..b5ad196ed 100644
--- a/modules/processing_vae.py
+++ b/modules/processing_vae.py
@@ -1,5 +1,6 @@
 import os
 import time
+import numpy as np
 import torch
 import torchvision.transforms.functional as TF
 from modules import shared, devices, sd_models, sd_vae, sd_vae_taesd
@@ -157,12 +158,15 @@ def vae_decode(latents, model, output_type='np', full_quality=True, width=None,
     else:
         decoded = taesd_vae_decode(latents=latents)
 
-    if hasattr(model, 'image_processor'):
-        imgs = model.image_processor.postprocess(decoded, output_type=output_type)
+    if torch.is_tensor(decoded):
+        if hasattr(model, 'image_processor'):
+            imgs = model.image_processor.postprocess(decoded, output_type=output_type)
+        else:
+            import diffusers
+            model.image_processor = diffusers.image_processor.VaeImageProcessor()
+            imgs = model.image_processor.postprocess(decoded, output_type=output_type)
     else:
-        import diffusers
-        model.image_processor = diffusers.image_processor.VaeImageProcessor()
-        imgs = model.image_processor.postprocess(decoded, output_type=output_type)
+        imgs = decoded if isinstance(decoded, list) or isinstance(decoded, np.ndarray) else [decoded]
 
     shared.state.job = prev_job
     if shared.cmd_opts.profile or debug:
diff --git a/modules/sd_samplers.py b/modules/sd_samplers.py
index b7c90e603..5694dd096 100644
--- a/modules/sd_samplers.py
+++ b/modules/sd_samplers.py
@@ -74,6 +74,9 @@ def create_sampler(name, model):
                 shared.log.warning(f'FLUX: sampler="{name}" unsupported')
                 # sampler.sampler.register_to_config(base_image_seq_len=256, max_image_seq_len=4096, base_shift=0.5, max_shift=1.15)
                 return None
+        if 'Lumina' in model.__class__.__name__:
+            shared.log.warning(f'AlphaVLLM-Lumina: sampler="{name}" unsupported')
+            return
         if not hasattr(model, 'scheduler_config'):
             model.scheduler_config = sampler.sampler.config.copy()
         model.scheduler = sampler.sampler
diff --git a/modules/sd_samplers_common.py b/modules/sd_samplers_common.py
index 1d4db96cb..1b1cd189a 100644
--- a/modules/sd_samplers_common.py
+++ b/modules/sd_samplers_common.py
@@ -41,10 +41,11 @@ def single_sample_to_image(sample, approximation=None):
                 approximation = 0
         # normal sample is [4,64,64]
         try:
-            if sample.dtype == torch.bfloat16:
+            if sample.dtype == torch.bfloat16 and (approximation == 0 or approximation == 1):
                 sample = sample.to(torch.float16)
         except Exception as e:
             warn_once(f'live preview: {e}')
+
         if len(sample.shape) > 4: # likely unknown video latent (e.g. svd)
             return Image.new(mode="RGB", size=(512, 512))
         if len(sample) == 16: # sd_cascade
@@ -58,6 +59,7 @@ def single_sample_to_image(sample, approximation=None):
             sample_min = torch.min(sample)
             if sample_min < -5:
                 sample = sample * (5 / abs(sample_min))
+
         if approximation == 2: # TAESD
             x_sample = sd_vae_taesd.decode(sample)
             x_sample = (1.0 + x_sample) / 2.0 # preview requires smaller range