Instructions to use Wan-AI/Wan2.2-Animate-14B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use Wan-AI/Wan2.2-Animate-14B with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("Wan-AI/Wan2.2-Animate-14B", dtype=torch.bfloat16, device_map="cuda") prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k" image = pipe(prompt).images[0] - Notebooks
- Google Colab
- Kaggle
RuntimeWarning: invalid value encountered in cast
I am running this work flow in runpod.io. But at the end of work flow the video generated is corrupted, there is black colored video with just voice.
I tried everything, even suggestions from Claude Opus 4.7. But still same error.
Anyone has faced same issue.
Below is the models used:
π ComfyUI/
βββ π models/
β βββ π clip/
β β βββ umt5-xxl-encoder-Q8_0.gguf
β βββ π clip_vision/
β β βββ clip_vision_h.safetensors
β βββ π loras/
β β βββ Wan2.2-I2V-A14B-lora-high_noise.safetensors
β β βββ Wan2.2-I2V-A14B-lora-low_noise.safetensors
β β βββ wan2.2_animate_14B_relight_lora_bf16.safetensors
β βββ π unet/
β β βββ Wan2.2-Animate-14B-Q8_0.gguf
β βββ π vae/
β βββ wan2.1_vae.safetensors
In the console log I could see below lines:
[INFO] Requested to load CLIPVisionModelProjection
[INFO] Model CLIPVisionModelProjection prepared for dynamic VRAM loading. 1205MB Staged. 0 patches attached. Force pre-loaded 132 weights: 330 KB.
[INFO] gguf qtypes: Q8_0 (169), F32 (73)
[INFO] Attempting to recreate sentencepiece tokenizer from GGUF file metadata...
[INFO] Created tokenizer with vocab size of 256384
[WARNING] Dequantizing token_embd.weight to prevent runtime OOM.
[INFO] CLIP/text encoder model load device: cuda:0, offload device: cpu, current: cpu, dtype: torch.float16
[INFO] Requested to load WanTEModel
[INFO] loaded completely; 19265.79 MB usable, 6695.95 MB loaded, full load: True
[INFO] Requested to load WanVAE
[INFO] Model WanVAE prepared for dynamic VRAM loading. 241MB Staged. 0 patches attached. Force pre-loaded 60 weights: 61 KB.
[INFO] Model WanVAE prepared for dynamic VRAM loading. 241MB Staged. 0 patches attached. Force pre-loaded 60 weights: 61 KB.
[INFO] Model WanVAE prepared for dynamic VRAM loading. 241MB Staged. 0 patches attached. Force pre-loaded 60 weights: 61 KB.
[INFO] gguf qtypes: F32 (929), Q8_0 (504), BF16 (8)
[INFO] model weight dtype torch.float16, manual cast: None
[INFO] model_type FLOW
[INFO] Requested to load WAN22_Animate
[INFO] loaded partially; 12215.27 MB usable, 12082.23 MB loaded, 5978.24 MB offloaded, 186.04 MB buffer reserved, lowvram patches: 0
[INFO] Attempting to release mmap (44)
100%|ββββββββββ| 10/10 [08:57<00:00, 53.71s/it]
[INFO] Requested to load WanVAE
[INFO] Model WanVAE prepared for dynamic VRAM loading. 241MB Staged. 0 patches attached. Force pre-loaded 60 weights: 61 KB.
/workspace/runpod-slim/ComfyUI/custom_nodes/comfyui-videohelpersuite/videohelpersuite/nodes.py:130: RuntimeWarning: invalid value encountered in cast
return tensor_to_int(tensor, 8).astype(np.uint8)
[INFO] Prompt executed in 00:14:27