File size: 1,662 Bytes
52abaef 7a7ba28 52abaef 32d1201 52abaef 7a7ba28 8b04b55 7a7ba28 8b04b55 4031025 8b04b55 52abaef 8b04b55 52abaef 8b04b55 52abaef 8b04b55 |
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 |
from typing import Dict, List, Any
import torch
from torch import autocast
from diffusers import StableDiffusionPipeline
import base64
from io import BytesIO
# set device
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
if device.type != 'cuda':
raise ValueError("need to run on GPU")
class EndpointHandler():
def __init__(self, path=""):
# load the optimized model
print("Loading model from:", path)
self.pipe = StableDiffusionPipeline.from_pretrained(path, torch_dtype=torch.float16)
self.pipe = self.pipe.to(device)
self.pipe.safety_checker = lambda images, clip_input: (images, None)
def __call__(self, data: Any) -> List[List[Dict[str, float]]]:
"""
Args:
data (:obj:):
includes the input data and the parameters for the inference.
Return:
A :obj:`dict`:. base64 encoded image
"""
inputs = data.pop("inputs", data)
print("Running inference with data:", data)
print("Running inference with inputs:", inputs)
out = self.pipe(
inputs,
guidance_scale = 5,
num_images_per_prompt = 1,
)
return out.images[0];
# run inference pipeline
# with autocast(device.type):
# image = self.pipe(inputs, guidance_scale = 5).images[0]
# # encode image as base 64
# buffered = BytesIO()
# image.save(buffered, format="JPEG")
# img_str = base64.b64encode(buffered.getvalue())
# postprocess the prediction
# return {"image": img_str.decode()} |