fix bugs

README.md

@@ -6,6 +6,8 @@ ComfyUI DynamiCrafter
 
 base workflow
 
-<img src="assets/wf_basic.png" raw=true>
+<img src="wf_basic.png" raw=true>
 
-https://github.com/chaojie/ComfyUI-DynamiCrafter/blob/main/workflow.json
+https://github.com/chaojie/ComfyUI-DynamiCrafter/blob/main/workflow.json
+
+<img src="video.gif" raw=true>

lvdm/models/autoencoder.py

@@ -6,8 +6,8 @@ from einops import rearrange
 import torch.nn.functional as F
 import pytorch_lightning as pl
 from ..modules.networks.ae_modules import Encoder, Decoder
-from .distributions import DiagonalGaussianDistribution
-from ...utils import instantiate_from_config
+from ..distributions import DiagonalGaussianDistribution
+from ...utils.utils import instantiate_from_config
 
 
 class AutoencoderKL(pl.LightningModule):

lvdm/modules/attention.py

@@ -9,12 +9,12 @@ try:
     XFORMERS_IS_AVAILBLE = True
 except:
     XFORMERS_IS_AVAILBLE = False
-from lvdm.common import (
+from ..common import (
     checkpoint,
     exists,
     default,
 )
-from lvdm.basics import zero_module
+from ..basics import zero_module
 
 
 class RelativePosition(nn.Module):

lvdm/modules/encoders/condition.py

@@ -4,8 +4,8 @@ import kornia
 import open_clip
 from torch.utils.checkpoint import checkpoint
 from transformers import T5Tokenizer, T5EncoderModel, CLIPTokenizer, CLIPTextModel
-from lvdm.common import autocast
-from utils.utils import count_params
+from ...common import autocast
+from ....utils.utils import count_params
 
 
 class AbstractEncoder(nn.Module):

lvdm/modules/networks/ae_modules.py

@@ -4,8 +4,8 @@ import torch
 import numpy as np
 import torch.nn as nn
 from einops import rearrange
-from utils.utils import instantiate_from_config
-from lvdm.modules.attention import LinearAttention
+from ....utils.utils import instantiate_from_config
+from ..attention import LinearAttention
 
 def nonlinearity(x):
     # swish

lvdm/modules/networks/openaimodel3d.py

@@ -4,16 +4,16 @@ import torch
 import torch.nn as nn
 from einops import rearrange
 import torch.nn.functional as F
-from lvdm.models.utils_diffusion import timestep_embedding
-from lvdm.common import checkpoint
-from lvdm.basics import (
+from ...models.utils_diffusion import timestep_embedding
+from ...common import checkpoint
+from ...basics import (
     zero_module,
     conv_nd,
     linear,
     avg_pool_nd,
     normalization
 )
-from lvdm.modules.attention import SpatialTransformer, TemporalTransformer
+from ..attention import SpatialTransformer, TemporalTransformer
 
 
 class TimestepBlock(nn.Module):

nodes.py

@@ -48,7 +48,7 @@ class DynamiCrafterSimple:
         image = 255.0 * image[0].cpu().numpy()
         #image = Image.fromarray(np.clip(image, 0, 255).astype(np.uint8))
         
-        imgs= image2video.get_image(image, prompt, steps, cfg_scale, eta, motion, seed)
+        imgs= model.get_image(image, prompt, steps, cfg_scale, eta, motion, seed)
         return imgs