Update webui.py

webui.py

@@ -24,6 +24,7 @@ from huggingface_hub import hf_hub_download
 
 dir = os.getcwd()
 
+@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
@@ -57,15 +58,15 @@ def list_models(name):
 def render_md(text):
     return f"{text}"
 
-def download_model(repo_id, filename):
-    hf_hub_download(
-        repo_id=repo_id,
-        filename=filename,
-        local_dir="models",
-        force_download=True, resume_download=False,
-        cache_dir=".cache",
-    )
-    return f"Downloaded!"
+
+hf_hub_download(
+    repo_id="ehristoforu/LLMs",
+    filename="llama-2-7b-chat.ggmlv3.q2_K.bin",
+    local_dir="models",
+    force_download=True, resume_download=False,
+    cache_dir=".cache",
+)
+
 
 history = []
 
@@ -76,7 +77,7 @@ If a question does not make any sense, or is not factually coherent, explain why
 """
 '''
 
-
+@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 "
@@ -149,11 +150,11 @@ with gr.Blocks(theme="theme-repo/STONE_Theme", title="TensorLM", css="style.css"
             
 
         with gr.Tab("💽"):
-            gr.Markdown("## Download model from 🤗 HuggingFace.co")
+            gr.Markdown("""## Download model from 🤗 HuggingFace.co (DON'T WORK ON HF.space)""")
             with gr.Row():
                 repo_id = gr.Textbox(label="REPO_ID",  value="ehristoforu/LLMs", lines=1, max_lines=1, interactive=False)
-                filename = gr.Dropdown(label="FILENAME", interactive=True, choices=["llama-2-7b-chat.ggmlv3.q2_K.bin", "llama-2-13b-chat.ggmlv3.q2_K.bin", "codellama-7b-instruct.ggmlv3.Q2_K.bin", "codellama-13b-instruct.ggmlv3.Q2_K.bin", "saiga-13b.ggmlv3.Q4_1.bin", "saiga-30b.ggmlv3.Q3_K.bin"], value="", allow_custom_value=False)
-                download_btn = gr.Button(value="Download")
+                filename = gr.Dropdown(label="FILENAME", interactive=False, choices=["llama-2-7b-chat.ggmlv3.q2_K.bin", "llama-2-13b-chat.ggmlv3.q2_K.bin", "codellama-7b-instruct.ggmlv3.Q2_K.bin", "codellama-13b-instruct.ggmlv3.Q2_K.bin", "saiga-13b.ggmlv3.Q4_1.bin", "saiga-30b.ggmlv3.Q3_K.bin"], value="", allow_custom_value=False)
+                download_btn = gr.Button(value="Download", interactive=False)
                 logs=gr.Markdown()
         with gr.Tab("📒"):
             with gr.Tab("Notebook"):
@@ -165,6 +166,7 @@ with gr.Blocks(theme="theme-repo/STONE_Theme", title="TensorLM", css="style.css"
                     markdown = gr.Markdown()
 
         with gr.Tab("⚙️"):
+            gr.Markdown("## Settings is DON'T WORK ON HF.space.")
             with gr.Row():
                 with gr.Column():
                     #with gr.Row():
@@ -172,7 +174,7 @@ with gr.Blocks(theme="theme-repo/STONE_Theme", title="TensorLM", css="style.css"
                     #    chat_style = gr.Dropdown(label="Style of chat", choices=["bubble", "panel"], value="bubble", interactive=True, allow_custom_value=False)
                     with gr.Row():
                         gr.Markdown("### Engine")
-                        reload_model = gr.Button("Apply settings to model", interactive=True)
+                        reload_model = gr.Button("Apply settings to model", interactive=False)
                         n_ctx = gr.Slider(label="Number of CTX", minimum=1024, maximum=4056, value=2048, step=8, interactive=True)
                         n_gpu_layers = gr.Slider(label="Number of GPU layers", minimum=0, maximum=36, value=0, step=1, interactive=True)
                         n_threads = gr.Slider(label="Number of Threads", minimum=2, maximum=36, value=4, step=1, interactive=True)
@@ -220,8 +222,7 @@ with gr.Blocks(theme="theme-repo/STONE_Theme", title="TensorLM", css="style.css"
 
 
 demo.launch(
-    inbrowser=True,
-    server_port=5555,
+    inbrowser=False,
     debug=False,
     quiet=True,
     favicon_path="assets/favicon.png",