Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -42,6 +42,8 @@ def initialize_database(list_file_obj):
|
|
42 |
def initialize_llmchain(llm_model, temperature, max_tokens, top_k, vector_db):
|
43 |
if vector_db is None:
|
44 |
return None, "Fehler: Keine Vektordatenbank verfügbar."
|
|
|
|
|
45 |
llm = HuggingFaceEndpoint(
|
46 |
repo_id=llm_model,
|
47 |
huggingfacehub_api_token=api_token,
|
@@ -89,7 +91,7 @@ def demo():
|
|
89 |
|
90 |
llm_btn = gr.Radio(["Flan-T5 Small", "Flan-T5 Base"], label="Verfügbare LLMs", value="Flan-T5 Small", type="index")
|
91 |
slider_temperature = gr.Slider(0.01, 1.0, 0.5, label="Temperature")
|
92 |
-
slider_maxtokens = gr.Slider(
|
93 |
slider_topk = gr.Slider(1, 10, 3, label="Top-k")
|
94 |
qachain_btn = gr.Button("Initialisiere QA-Chatbot")
|
95 |
llm_status = gr.Textbox(label="Chatbot-Status", value="Nicht initialisiert", interactive=False)
|
|
|
42 |
def initialize_llmchain(llm_model, temperature, max_tokens, top_k, vector_db):
|
43 |
if vector_db is None:
|
44 |
return None, "Fehler: Keine Vektordatenbank verfügbar."
|
45 |
+
if max_tokens > 250:
|
46 |
+
max_tokens = 250 # Begrenze max_new_tokens, um Fehler zu vermeiden
|
47 |
llm = HuggingFaceEndpoint(
|
48 |
repo_id=llm_model,
|
49 |
huggingfacehub_api_token=api_token,
|
|
|
91 |
|
92 |
llm_btn = gr.Radio(["Flan-T5 Small", "Flan-T5 Base"], label="Verfügbare LLMs", value="Flan-T5 Small", type="index")
|
93 |
slider_temperature = gr.Slider(0.01, 1.0, 0.5, label="Temperature")
|
94 |
+
slider_maxtokens = gr.Slider(1, 250, 128, label="Max Tokens") # Begrenzung auf 250
|
95 |
slider_topk = gr.Slider(1, 10, 3, label="Top-k")
|
96 |
qachain_btn = gr.Button("Initialisiere QA-Chatbot")
|
97 |
llm_status = gr.Textbox(label="Chatbot-Status", value="Nicht initialisiert", interactive=False)
|