wip

nodes.py

@@ -1,9 +1,12 @@
 from .svd import load_model, get_unique_embedder_keys_from_conditioner, get_batch
+from torchvision.transforms import ToTensor, ToPILImage
+from einops import rearrange, repeat
 import gc
 import folder_paths
 import torch
 import os
 import math
+import numpy as np
 
 class SVDModelLoader:
     def __init__(self):
@@ -84,9 +87,6 @@ class SVDSampler:
                 "seed" : ("INT", {
                     "default": 23,
                 }),
-                "decoding_t" : ("INT", {
-                    "default": 14,
-                }),
                 "device" : (devices,),
             },
         }
@@ -100,31 +100,37 @@ class SVDSampler:
 
     CATEGORY = "Comfy Stable Video Diffusion"
 
-    def sample_video(self, image, model, motion_bucket_id, fps_id, cond_aug, seed, decoding_t, device):
-        # convert image tensor to PIL image
-        print(type(image))
-        print(image.shape)
-        1/0
+    def sample_video(self, image, model, motion_bucket_id, fps_id, cond_aug, seed, device):
+        # convert image torch tensor to PIL image
+        # image shape: (1, H, W, C)
+        image = image.squeeze(0)
+        image = image.permute(2, 0, 1)
+        image = (image + 1.0) / 2.0
+        image = image.clamp(min=0.0, max=1.0)
+
+        image = ToPILImage()(image)
 
         if image.mode == "RGBA":
             image = image.convert("RGB")
-            w, h = image.size
+        
+        w, h = image.size
 
-            if h % 64 != 0 or w % 64 != 0:
-                width, height = map(lambda x: x - x % 64, (w, h))
-                image = image.resize((width, height))
-                print(
-                    f"WARNING: Your image is of size {h}x{w} which is not divisible by 64. We are resizing to {height}x{width}!"
-                )
+        if h % 64 != 0 or w % 64 != 0:
+            width, height = map(lambda x: x - x % 64, (w, h))
+            image = image.resize((width, height))
+            print(
+                f"WARNING: Your image is of size {h}x{w} which is not divisible by 64. We are resizing to {height}x{width}!"
+            )
 
-            image = ToTensor()(image)
-            image = image * 2.0 - 1.0
+        image = ToTensor()(image)
+        image = image * 2.0 - 1.0
 
         image = image.unsqueeze(0).to(device)
         H, W = image.shape[2:]
         assert image.shape[1] == 3
         F = 8
         C = 4
+        num_frames = model.sampler.guider.num_frames
         shape = (num_frames, C, H // F, W // F)
         if (H, W) != (576, 1024):
             print(
@@ -149,8 +155,11 @@ class SVDSampler:
         value_dict["cond_frames"] = image + cond_aug * torch.randn_like(image)
         value_dict["cond_aug"] = cond_aug
 
+        all_samples_z = []
+
         with torch.no_grad():
             with torch.autocast(device):
+                print("getting batch")
                 batch, batch_uc = get_batch(
                     get_unique_embedder_keys_from_conditioner(model.conditioner),
                     value_dict,
@@ -158,6 +167,8 @@ class SVDSampler:
                     T=num_frames,
                     device=device,
                 )
+
+                print("getting conditioning")
                 c, uc = model.conditioner.get_unconditional_conditioning(
                     batch,
                     batch_uc=batch_uc,
@@ -167,6 +178,7 @@ class SVDSampler:
                     ],
                 )
 
+                print("repeating conditioning")
                 for k in ["crossattn", "concat"]:
                     uc[k] = repeat(uc[k], "b ... -> b t ...", t=num_frames)
                     uc[k] = rearrange(uc[k], "b t ... -> (b t) ...", t=num_frames)
@@ -186,7 +198,10 @@ class SVDSampler:
                         model.model, input, sigma, c, **additional_model_inputs
                     )
 
+                print("sampling: ", randn.shape)
                 samples_z = model.sampler(denoiser, randn, cond=c, uc=uc)
+                print("sampled: ", samples_z.shape)
+        
         return (samples_z,)
 
 
@@ -215,12 +230,20 @@ class SVDDecoder:
     CATEGORY = "Comfy Stable Video Diffusion"
 
     def decode(self, samples_z, model, decoding_t, device):
+        print("decoding: ", samples_z.shape)
         with torch.no_grad():
             with torch.autocast(device):
                 model.en_and_decode_n_samples_a_time = decoding_t
                 samples_x = model.decode_first_stage(samples_z)
                 samples = torch.clamp((samples_x + 1.0) / 2.0, min=0.0, max=1.0)
-        return (samples,)
+
+                vid = (
+                    (rearrange(samples, "t c h w -> t h w c") * 255)
+                    .cpu()
+                    .numpy()
+                    .astype(np.uint8)
+                )
+        return (vid,)
 
 
 # A dictionary that contains all nodes you want to export with their names