Update webui.py

webui.py

@@ -17,7 +17,7 @@ import llama_cpp
 from llama_cpp import Llama
 import random
 from huggingface_hub import hf_hub_download
-import spaces
+#import spaces
 
 #from blip.blip_engine import blip_run
 
@@ -25,7 +25,7 @@ import spaces
 
 dir = os.getcwd()
 
-@spaces.GPU
+#@spaces.GPU
 def load_model(path, n_ctx, n_gpu_layers, n_threads, verbose, f16_kv, logits_all, vocab_only, use_mmap, use_mlock, n_batch, last_n_tokens_size, low_vram, rope_freq_base, rope_freq_scale):
     try:
         global llm
@@ -78,7 +78,7 @@ If a question does not make any sense, or is not factually coherent, explain why
 """
 '''
 
-@spaces.GPU
+#@spaces.GPU
 def generate_text(message, history, system_prompt, preset, temperature, max_tokens, top_p, top_k, repeat_penalty):
     temp = ""
     input_prompt = f"[INST] <<SYS>>\nYou are {preset}. {system_prompt}.\n<</SYS>>\n\n "
@@ -217,7 +217,7 @@ with gr.Blocks(theme="theme-repo/STONE_Theme", title="TensorLM", css="style.css"
 
     
                 
-    model.change(load_model, inputs=[model, n_ctx, n_gpu_layers, n_threads, verbose, f16_kv, logits_all, vocab_only, use_mmap, use_mlock, n_batch, last_n_tokens_size, low_vram, rope_freq_base, rope_freq_scale], outputs=model, api_name=False, queue=True)
+    model.change(load_model, inputs=[model, n_ctx, n_gpu_layers, n_threads, verbose, f16_kv, logits_all, vocab_only, use_mmap, use_mlock, n_batch, last_n_tokens_size, low_vram, rope_freq_base, rope_freq_scale], outputs=model, api_name=False, queue=False)
     reload_model.click(load_model, inputs=[model, n_ctx, n_gpu_layers, n_threads, verbose, f16_kv, logits_all, vocab_only, use_mmap, use_mlock, n_batch, last_n_tokens_size, low_vram, rope_freq_base, rope_freq_scale], outputs=model, api_name=False, queue=True)