Update

app.py

@@ -10,26 +10,26 @@ import torch
 # }
 
 model_names = {
-    "LeoLM_13B": "LeoLM/leo-hessianai-7b",
+    "LeoLM_7B": "LeoLM/leo-hessianai-7b",
     "Occiglot_7B": "occiglot/occiglot-7b-de-en"
 }
 
 # Tokenizer und Modelle vorbereiten
 tokenizers = {name: AutoTokenizer.from_pretrained(model) for name, model in model_names.items()}
-models = {name: AutoModelForCausalLM.from_pretrained(model, device_map="auto", torch_dtype=torch.float16) for name, model in model_names.items()}
+models = {name: AutoModelForCausalLM.from_pretrained(model) for name, model in model_names.items()}
 
 # Funktion zur Generierung der Antwort
 def generate_response(model_choice, prompt):
     tokenizer = tokenizers[model_choice]
     model = models[model_choice]
-    inputs = tokenizer(prompt, return_tensors="pt").to("cuda")
+    inputs = tokenizer(prompt, return_tensors="pt")
     outputs = model.generate(inputs["input_ids"], max_new_tokens=100, do_sample=True)
     response = tokenizer.decode(outputs[0], skip_special_tokens=True)
     return response
 
 # Gradio Interface
 with gr.Blocks() as demo:
-    gr.Markdown("# Vergleich von LLMs: LeoLM, Occiglot und LLaMA 2")
+    gr.Markdown("# Vergleich von LLMs: LeoLM und Occiglot")
     with gr.Row():
         model_choice = gr.Radio(list(model_names.keys()), label="Modell auswählen")
         prompt = gr.Textbox(label="Frage stellen", placeholder="Was sind die Hauptursachen für Bluthochdruck?")