Update app.py

app.py

@@ -1,4 +1,3 @@
-import deepspeed
 import torch
 from transformers import pipeline
 import os
@@ -24,21 +23,6 @@ device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
 print('Using device:', device)
 print()
 
-total_mem = 0
-if device.type == 'cuda':
-    print(torch.cuda.get_device_name(0))
-    total_mem = round(torch.cuda.get_device_properties(0).total_memory / 1024**3, 1)
-    print('Total Memory: ', total_mem, 'GB')
-
-should_replace_with_kernel_inject = total_mem >= 12
-print(f'should_replace_with_kernel_inject = {should_replace_with_kernel_inject}')
-
-ds_engine = deepspeed.init_inference(generator.model,
-                                     mp_size=world_size,
-                                     dtype=torch.half,
-                                     replace_with_kernel_inject=should_replace_with_kernel_inject)
-generator.model = ds_engine.module
-
 # פונקציית יצירת הטקסט
 def chat_with_model(history):
     prompt = history[-1]["content"]