Update app.py

app.py

@@ -9,19 +9,20 @@ import requests
 import gradio as gr
 import numpy as np
 from lxml.html import fromstring
-from transformers import pipeline
+#from transformers import pipeline
 from torch import multiprocessing as mp, nn
 #from torch.multiprocessing import Pool
 #from pathos.multiprocessing import ProcessPool as Pool
-from pathos.threading import ThreadPool as Pool
-from diffusers.pipelines.flux import FluxPipeline
+#from pathos.threading import ThreadPool as Pool
+#from diffusers.pipelines.flux import FluxPipeline
 from diffusers.utils import export_to_gif, load_image
 from diffusers.models.modeling_utils import ModelMixin
 from huggingface_hub import hf_hub_download
 from safetensors.torch import load_file, save_file
 from diffusers import DiffusionPipeline, AnimateDiffPipeline, MotionAdapter, EulerDiscreteScheduler, DDIMScheduler, StableDiffusionXLPipeline, UNet2DConditionModel, AutoencoderKL, UNet3DConditionModel
-import jax
-import jax.numpy as jnp
+#import jax
+#import jax.numpy as jnp
+from pyina.launchers import TorqueMpiPool as Pool
 import sys
 import warnings
 
@@ -210,17 +211,11 @@ def run(i,m,p1,p2,*result):
         p2_en = translate(p2,"english")
         pm = {"p":p1_en,"n":p2_en,"m":m,"i":i}
         ln = len(result)
-        print("Threads: "+str(ln))
         rng = list(range(ln))
-    
         arr = [pm for _ in rng]
-        pool = Pool(ln)
-        out = list(pool.imap(infer,arr))
-        pool.close()
-        pool.join()
-        pool.clear()
 
-        return out
+        with Pool(f'{ ln }:ppn=2', queue='productionQ', timelimit='5:00:00', workdir='.') as pool:
+            return pool.map(infer,arr)
 
 pipe = AnimateDiffPipeline.from_pretrained(base, motion_adapter=adapter, torch_dtype=dtype).to(device)
 pipe.scheduler = DDIMScheduler(