Update chatbot.py

chatbot.py

@@ -5,11 +5,11 @@ import random
 from threading import Thread
 from typing import List, Dict, Union
 import subprocess
-subprocess.run(
-    "pip install flash-attn --no-build-isolation",
-    env={"FLASH_ATTENTION_SKIP_CUDA_BUILD": "TRUE"},
-    shell=True,
-)
+# subprocess.run(
+#     "pip install flash-attn --no-build-isolation",
+#     env={"FLASH_ATTENTION_SKIP_CUDA_BUILD": "TRUE"},
+#     shell=True,
+# )
 import torch
 import gradio as gr
 from bs4 import BeautifulSoup
@@ -31,7 +31,7 @@ model_id = "llava-hf/llava-interleave-qwen-7b-hf"
 # model_id = "llava-hf/llava-interleave-qwen-7b-dpo-hf"
 processor = LlavaProcessor.from_pretrained(model_id)
 model = LlavaForConditionalGeneration.from_pretrained(model_id,torch_dtype=torch.float16,  use_flash_attention_2=True)
-model.to("cuda")
+model.to("cpu")
 # Credit to merve for code of llava interleave qwen
 
 GROQ_API_KEY = os.environ.get("GROQ_API_KEY", None)