File size: 2,182 Bytes
c9fea2f 3930573 4419d94 9b13d16 4419d94 9b13d16 c9fea2f 9b13d16 4419d94 9b13d16 c9fea2f 9b13d16 c9fea2f 9b13d16 c9fea2f 9b13d16 c9fea2f 9b13d16 c9fea2f 9b13d16 4419d94 0626129 c9fea2f |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 |
import os, tempfile, json, traceback
import gradio as gr
from huggingface_hub import InferenceClient
MODEL_ID = "Wan-AI/Wan2.2-T2V-A14B-Diffusers"
def _pick_token(request: gr.Request | None):
tok = os.getenv("HF_TOKEN") # Space secret first
if not tok and request:
auth = request.headers.get("authorization", "")
if isinstance(auth, str) and auth.lower().startswith("bearer "):
tok = auth.split(" ", 1)[1]
tok = tok or request.headers.get("x-hf-token")
return tok
def t2v(prompt: str, fps: int = 12, request: gr.Request | None = None):
try:
token = _pick_token(request)
if not token:
raise RuntimeError("No token found. Set Space secret HF_TOKEN or send Authorization: Bearer hf_...")
client = InferenceClient(provider="fal-ai", api_key=token)
video_bytes = client.text_to_video(prompt, model=MODEL_ID)
if hasattr(video_bytes, "read"):
video_bytes = video_bytes.read()
if not isinstance(video_bytes, (bytes, bytearray)):
raise RuntimeError(f"Unexpected provider return type: {type(video_bytes)}")
tmp = tempfile.NamedTemporaryFile(suffix=".mp4", delete=False)
tmp.write(video_bytes); tmp.flush(); tmp.close()
return tmp.name
except Exception as e:
err = {"error": str(e), "traceback": traceback.format_exc(), "has_secret": bool(os.getenv("HF_TOKEN")), "model": MODEL_ID}
raise gr.Error(json.dumps(err, ensure_ascii=False))
def health():
return {"ok": True, "model": MODEL_ID, "has_secret": bool(os.getenv("HF_TOKEN"))}
with gr.Blocks() as demo:
gr.Markdown("### Wan2.2 T2V via fal-ai · HTTP API at `/api/predict` and `/api/health`")
gr.Interface(
fn=t2v,
inputs=[gr.Textbox(label="Prompt"), gr.Slider(4, 24, value=12, step=1, label="FPS")],
outputs=gr.Video(label="Video"),
api_name="predict",
title="Wan2.2 Text-to-Video",
flagging_mode="never",
)
gr.Interface(fn=health, inputs=[], outputs="json", api_name="health")
if __name__ == "__main__":
demo.launch(ssr_mode=False) # keep SSR off to avoid Node shutdown loops
|