GenAI-Arena / model /model_manager.py
yuanshengni's picture
Update model/model_manager.py
5c6a38a verified
raw
history blame
5.97 kB
import concurrent.futures
import random
import gradio as gr
# from fal_api_utils import load_fal_model
from .imagenhub_utils import load_imagenhub_model
import spaces
import requests
import io, base64, json
from PIL import Image
import os
IMAGE_GENERATION_MODELS = ['imagenhub_LCM_generation','imagenhub_SDXLTurbo_generation','imagenhub_SDXL_generation', 'imagenhub_PixArtAlpha_generation',
'imagenhub_OpenJourney_generation','imagenhub_SDXLLightning_generation', 'imagenhub_StableCascade_generation',
'imagenhub_PlayGroundV2_generation', 'imagenhub_PlayGroundV2.5_generation']
IMAGE_EDITION_MODELS = ['imagenhub_CycleDiffusion_edition', 'imagenhub_Pix2PixZero_edition', 'imagenhub_Prompt2prompt_edition',
'imagenhub_SDEdit_edition', 'imagenhub_InstructPix2Pix_edition', 'imagenhub_MagicBrush_edition', 'imagenhub_PNP_edition']
class ModelManager:
def __init__(self):
self.model_ig_list = IMAGE_GENERATION_MODELS
self.model_ie_list = IMAGE_EDITION_MODELS
self.loaded_models = {}
# @spaces.GPU
def load_model_pipe(self, model_name):
model_source, model_name, model_type = model_name.split("_")
if not model_name in self.loaded_models:
if model_source == "imagenhub":
pipe = load_imagenhub_model(model_name)
# elif model_source == "fal":
# pipe = load_fal_model(model_name, model_type)
else:
raise ValueError(f"Model source {model_source} not supported")
self.loaded_models[model_name] = pipe
else:
pipe = self.loaded_models[model_name]
return pipe
def generate_image_playground(self, model_name, prompt):
if model_name == "imagenhub_PlayGroundV2_generation":
model_name = "Playground_v2"
elif model_name == "imagenhub_PlayGroundV2.5_generation":
model_name = "Playground_v2.5"
headers = {
'Content-Type': 'application/json',
'Authorization': os.environ['PlaygroundAPI'],
}
data = json.dumps({"prompt": prompt, "filter_model": model_name, "scheduler": "DPMPP_2M_K", "guidance_scale": 3})
response = requests.post('https://playground.com/api/models/external/v1', headers=headers, data=data)
response.raise_for_status()
json_obj = response.json()
image_base64 = json_obj['images'][0]
img = Image.open(io.BytesIO(base64.decodebytes(bytes(image_base64, "utf-8"))))
return img
@spaces.GPU(duration=60)
def generate_image_ig(self, prompt, model_name):
if 'playground' in model_name.lower():
result = self.generate_image_playground(model_name=model_name, prompt=prompt)
else:
pipe = self.load_model_pipe(model_name)
result = pipe(prompt=prompt)
return result
def generate_image_ig_parallel_anony(self, prompt, model_A, model_B):
if model_A == "" and model_B == "":
model_names = random.sample([model for model in self.model_ig_list], 2)
else:
model_names = [model_A, model_B]
results = []
with concurrent.futures.ThreadPoolExecutor() as executor:
future_to_result = {executor.submit(self.generate_image_ig, prompt, model): model for model in model_names}
for future in concurrent.futures.as_completed(future_to_result):
result = future.result()
results.append(result)
return results[0], results[1], model_names[0], model_names[1]
def generate_image_ig_parallel(self, prompt, model_A, model_B):
results = []
model_names = [model_A, model_B]
with concurrent.futures.ThreadPoolExecutor() as executor:
future_to_result = {executor.submit(self.generate_image_ig, prompt, model): model for model in model_names}
for future in concurrent.futures.as_completed(future_to_result):
result = future.result()
results.append(result)
return results[0], results[1]
@spaces.GPU(duration=150)
def generate_image_ie(self, textbox_source, textbox_target, textbox_instruct, source_image, model_name):
pipe = self.load_model_pipe(model_name)
result = pipe(src_image = source_image, src_prompt = textbox_source, target_prompt = textbox_target, instruct_prompt = textbox_instruct)
return result
def generate_image_ie_parallel(self, textbox_source, textbox_target, textbox_instruct, source_image, model_A, model_B):
results = []
model_names = [model_A, model_B]
with concurrent.futures.ThreadPoolExecutor() as executor:
future_to_result = {executor.submit(self.generate_image_ie, textbox_source, textbox_target, textbox_instruct, source_image, model): model for model in model_names}
for future in concurrent.futures.as_completed(future_to_result):
result = future.result()
results.append(result)
return results[0], results[1]
def generate_image_ie_parallel_anony(self, textbox_source, textbox_target, textbox_instruct, source_image, model_A, model_B):
if model_A == "" and model_B == "":
model_names = random.sample([model for model in self.model_ie_list], 2)
else:
model_names = [model_A, model_B]
results = []
# model_names = [model_A, model_B]
with concurrent.futures.ThreadPoolExecutor() as executor:
future_to_result = {executor.submit(self.generate_image_ie, textbox_source, textbox_target, textbox_instruct, source_image, model): model for model in model_names}
for future in concurrent.futures.as_completed(future_to_result):
result = future.result()
results.append(result)
return results[0], results[1], model_names[0], model_names[1]