Rob1221rib/wan22-qx-encoders-gguf

Instalación y uso básico con Diffusers:

pip install -U diffusers transformers accelerate

import torch
from diffusers import DiffusionPipeline

# switch to "mps" for apple devices
pipe = DiffusionPipeline.from_pretrained(
    "Rob1221rib/wan22-qx-encoders-gguf",
    dtype=torch.bfloat16,
    device_map="cuda"
)

prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k"
image = pipe(prompt).images[0]

Carga directa de un encoder GGUF con llama-cpp-python:

# !pip install llama-cpp-python
from llama_cpp import Llama

llm = Llama.from_pretrained(
    repo_id="Rob1221rib/wan22-qx-encoders-gguf",
    filename="text_encoders/umt5-xxl-encoder-q3-k-m.gguf",
)

output = llm(
    "Once upon a time,",
    max_tokens=512,
    echo=True
)
print(output)

Uso local con llama.cpp:

curl -LsSf https://llama.app/install.sh | sh

# Start a local OpenAI-compatible server with a web UI:
llama serve -hf Rob1221rib/wan22-qx-encoders-gguf

# Run inference directly in the terminal:
llama cli -hf Rob1221rib/wan22-qx-encoders-gguf

Uso con Ollama o Docker Model Runner:

ollama run hf.co/Rob1221rib/wan22-qx-encoders-gguf

docker model run hf.co/Rob1221rib/wan22-qx-encoders-gguf

Ejemplo de selección de encoder en un pipeline WAN 2.2 QX:

from diffusers import DiffusionPipeline
import torch

pipe = DiffusionPipeline.from_pretrained(
    "path/to/wan-2.2-qx",
    text_encoder_path="E:/huggingface/wan22-qx-encoders-gguf/text_encoders/umt5-xxl-encoder-q4-k-m.gguf",
    torch_dtype=torch.float16,
    variant="fp16"
)

pipe = pipe.to("cuda")

prompt = "A serene mountain landscape at sunset with flowing clouds"
video_frames = pipe(
    prompt=prompt,
    num_frames=24,
    num_inference_steps=30
).frames

save_video(video_frames, "output.mp4", fps=8)