Spaces:
Paused
Paused
Update main.py
Browse files
main.py
CHANGED
|
@@ -1,16 +1,74 @@
|
|
| 1 |
-
import
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2 |
import AnimeGANv3_src
|
| 3 |
-
if __name__ == '__main__':
|
| 4 |
|
| 5 |
-
|
| 6 |
-
input_imgs_path = r'../../v3-usa\dataset\USA\val'
|
| 7 |
-
# input_imgs_path = r'/mnt/data/xinchen/v3-usa/dataset/USA/val'
|
| 8 |
-
output_path = 'AnimeGANv3_usa_64_output'
|
| 9 |
-
# img = cv2.imread(os.path.join(input_imgs_path, os.listdir(input_imgs_path)[0]))
|
| 10 |
-
img = cv2.imread(os.path.join(input_imgs_path, 'jp_16.png'))
|
| 11 |
-
out = AnimeGANv3_src.Convert(cv2.cvtColor(img, cv2.COLOR_BGR2RGB), f, True)
|
| 12 |
-
# cv2.imshow('d', cv2.cvtColor(out, cv2.COLOR_BGR2RGB))
|
| 13 |
-
# cv2.waitKey(0)
|
| 14 |
-
cv2.imwrite('a.jpg', cv2.cvtColor(out, cv2.COLOR_BGR2RGB))
|
| 15 |
|
|
|
|
| 16 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import os
|
| 2 |
+
import cv2
|
| 3 |
+
from fastapi import FastAPI, File, UploadFile
|
| 4 |
+
from fastapi import FastAPI, File, UploadFile, Form, Request
|
| 5 |
+
from fastapi.responses import HTMLResponse, FileResponse
|
| 6 |
+
from fastapi.staticfiles import StaticFiles
|
| 7 |
+
from fastapi.templating import Jinja2Templates
|
| 8 |
+
from fastapi import FastAPI, File, UploadFile, HTTPException
|
| 9 |
+
from fastapi.responses import JSONResponse
|
| 10 |
+
from fastapi.responses import StreamingResponse
|
| 11 |
+
from fastapi import FastAPI, File, UploadFile
|
| 12 |
+
from fastapi.responses import FileResponse
|
| 13 |
+
from pydantic import BaseModel
|
| 14 |
+
import shutil
|
| 15 |
import AnimeGANv3_src
|
|
|
|
| 16 |
|
| 17 |
+
app = FastAPI()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 18 |
|
| 19 |
+
os.makedirs('output', exist_ok=True)
|
| 20 |
|
| 21 |
+
class InferenceRequest(BaseModel):
|
| 22 |
+
img_path: str
|
| 23 |
+
Style: str
|
| 24 |
+
if_face: str
|
| 25 |
+
|
| 26 |
+
def inference(img_path, Style, if_face=None):
|
| 27 |
+
print(img_path, Style, if_face)
|
| 28 |
+
try:
|
| 29 |
+
img = cv2.imread(img_path)
|
| 30 |
+
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
|
| 31 |
+
if Style == "AnimeGANv3_Arcane":
|
| 32 |
+
f = "A"
|
| 33 |
+
elif Style == "AnimeGANv3_Trump v1.0":
|
| 34 |
+
f = "T"
|
| 35 |
+
elif Style == "AnimeGANv3_Shinkai":
|
| 36 |
+
f = "S"
|
| 37 |
+
elif Style == "AnimeGANv3_PortraitSketch":
|
| 38 |
+
f = "P"
|
| 39 |
+
elif Style == "AnimeGANv3_Hayao":
|
| 40 |
+
f = "H"
|
| 41 |
+
elif Style == "AnimeGANv3_Disney v1.0":
|
| 42 |
+
f = "D"
|
| 43 |
+
elif Style == "AnimeGANv3_JP_face v1.0":
|
| 44 |
+
f = "J"
|
| 45 |
+
elif Style == "AnimeGANv3_Kpop v2.0":
|
| 46 |
+
f = "K"
|
| 47 |
+
else:
|
| 48 |
+
f = "U"
|
| 49 |
+
|
| 50 |
+
try:
|
| 51 |
+
det_face = True if if_face=="Yes" else False
|
| 52 |
+
output = AnimeGANv3_src.Convert(img, f, det_face)
|
| 53 |
+
save_path = f"output/out.{img_path.rsplit('.')[-1]}"
|
| 54 |
+
cv2.imwrite(save_path, output[:, :, ::-1])
|
| 55 |
+
return output, save_path
|
| 56 |
+
except RuntimeError as error:
|
| 57 |
+
print('Error', error)
|
| 58 |
+
except Exception as error:
|
| 59 |
+
print('global exception', error)
|
| 60 |
+
return None, None
|
| 61 |
+
|
| 62 |
+
@app.post("/inference/")
|
| 63 |
+
async def inference_api(request: InferenceRequest):
|
| 64 |
+
img_path = request.img_path
|
| 65 |
+
Style = request.Style
|
| 66 |
+
if_face = request.if_face
|
| 67 |
+
output, save_path = inference(img_path, Style, if_face)
|
| 68 |
+
return {"output": output, "save_path": save_path}
|
| 69 |
+
|
| 70 |
+
app.mount("/", StaticFiles(directory="static", html=True), name="static")
|
| 71 |
+
|
| 72 |
+
@app.get("/")
|
| 73 |
+
def index() -> FileResponse:
|
| 74 |
+
return FileResponse(path="/app/static/index.html", media_type="text/html")
|