Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -1,24 +1,20 @@
|
|
| 1 |
import torch
|
| 2 |
-
import gradio as gr
|
| 3 |
from transformers import AutoTokenizer, AutoModelForCausalLM
|
|
|
|
| 4 |
|
| 5 |
-
# Optymalizacja
|
| 6 |
torch.set_float32_matmul_precision('high')
|
| 7 |
|
| 8 |
-
#
|
| 9 |
-
model_name = "
|
| 10 |
-
|
| 11 |
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
| 12 |
model = AutoModelForCausalLM.from_pretrained(model_name)
|
| 13 |
|
| 14 |
-
#
|
| 15 |
-
MAX_HISTORY = 800
|
| 16 |
|
| 17 |
-
# Funkcja czatu
|
| 18 |
def chatbot_response(user_input, history, top_k, top_p, temperature):
|
| 19 |
history += f"Użytkownik: {user_input}\nAI:"
|
| 20 |
input_ids = tokenizer.encode(history, return_tensors="pt", truncation=True, max_length=1024)
|
| 21 |
-
|
| 22 |
if input_ids.shape[1] > MAX_HISTORY:
|
| 23 |
input_ids = input_ids[:, -MAX_HISTORY:]
|
| 24 |
|
|
@@ -32,14 +28,13 @@ def chatbot_response(user_input, history, top_k, top_p, temperature):
|
|
| 32 |
temperature=temperature
|
| 33 |
)
|
| 34 |
|
| 35 |
-
|
| 36 |
-
|
| 37 |
-
history += f" {
|
| 38 |
return history, history
|
| 39 |
|
| 40 |
-
|
| 41 |
-
|
| 42 |
-
gr.Markdown("# 🤖 Polski Chatbot AI\nModel: distilgpt2-finetuned-text-generation")
|
| 43 |
|
| 44 |
chat_output = gr.Textbox(label="Historia rozmowy", lines=15, interactive=False)
|
| 45 |
user_input = gr.Textbox(label="Wpisz wiadomość")
|
|
@@ -51,7 +46,6 @@ with gr.Blocks(title="Polski Chatbot AI") as demo:
|
|
| 51 |
history_state = gr.State("")
|
| 52 |
|
| 53 |
send_btn = gr.Button("Wyślij")
|
| 54 |
-
|
| 55 |
send_btn.click(
|
| 56 |
chatbot_response,
|
| 57 |
inputs=[user_input, history_state, top_k, top_p, temperature],
|
|
@@ -61,11 +55,12 @@ with gr.Blocks(title="Polski Chatbot AI") as demo:
|
|
| 61 |
clear_btn = gr.Button("🧹 Wyczyść historię")
|
| 62 |
clear_btn.click(lambda: ("", ""), outputs=[chat_output, history_state])
|
| 63 |
|
| 64 |
-
gr.Markdown("
|
| 65 |
with gr.Row():
|
| 66 |
-
gr.Button("Jak się nazywasz?").click(
|
| 67 |
-
|
|
|
|
|
|
|
| 68 |
|
| 69 |
-
# Uruchom
|
| 70 |
if __name__ == "__main__":
|
| 71 |
demo.launch()
|
|
|
|
| 1 |
import torch
|
|
|
|
| 2 |
from transformers import AutoTokenizer, AutoModelForCausalLM
|
| 3 |
+
import gradio as gr
|
| 4 |
|
| 5 |
+
# Optymalizacja obliczeń
|
| 6 |
torch.set_float32_matmul_precision('high')
|
| 7 |
|
| 8 |
+
# Nowy, lżejszy model
|
| 9 |
+
model_name = "distilbert/distilgpt2"
|
|
|
|
| 10 |
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
| 11 |
model = AutoModelForCausalLM.from_pretrained(model_name)
|
| 12 |
|
| 13 |
+
MAX_HISTORY = 800 # limit tokenów w historii
|
|
|
|
| 14 |
|
|
|
|
| 15 |
def chatbot_response(user_input, history, top_k, top_p, temperature):
|
| 16 |
history += f"Użytkownik: {user_input}\nAI:"
|
| 17 |
input_ids = tokenizer.encode(history, return_tensors="pt", truncation=True, max_length=1024)
|
|
|
|
| 18 |
if input_ids.shape[1] > MAX_HISTORY:
|
| 19 |
input_ids = input_ids[:, -MAX_HISTORY:]
|
| 20 |
|
|
|
|
| 28 |
temperature=temperature
|
| 29 |
)
|
| 30 |
|
| 31 |
+
decoded = tokenizer.decode(output[0], skip_special_tokens=True)
|
| 32 |
+
reply = decoded[len(history):].split("Użytkownik:")[0].strip()
|
| 33 |
+
history += f" {reply}\n"
|
| 34 |
return history, history
|
| 35 |
|
| 36 |
+
with gr.Blocks() as demo:
|
| 37 |
+
gr.Markdown("# 🤖 Polski Chatbot AI (DistilGPT2)")
|
|
|
|
| 38 |
|
| 39 |
chat_output = gr.Textbox(label="Historia rozmowy", lines=15, interactive=False)
|
| 40 |
user_input = gr.Textbox(label="Wpisz wiadomość")
|
|
|
|
| 46 |
history_state = gr.State("")
|
| 47 |
|
| 48 |
send_btn = gr.Button("Wyślij")
|
|
|
|
| 49 |
send_btn.click(
|
| 50 |
chatbot_response,
|
| 51 |
inputs=[user_input, history_state, top_k, top_p, temperature],
|
|
|
|
| 55 |
clear_btn = gr.Button("🧹 Wyczyść historię")
|
| 56 |
clear_btn.click(lambda: ("", ""), outputs=[chat_output, history_state])
|
| 57 |
|
| 58 |
+
gr.Markdown("## 🔄 Szybkie pytania:")
|
| 59 |
with gr.Row():
|
| 60 |
+
gr.Button("Jak się nazywasz?").click(
|
| 61 |
+
lambda _: chatbot_response("Jak się nazywasz?", "", 50, 0.9, 0.7), outputs=[chat_output, history_state])
|
| 62 |
+
gr.Button("Czym się zajmujesz?").click(
|
| 63 |
+
lambda _: chatbot_response("Czym się zajmujesz?", "", 50, 0.9, 0.7), outputs=[chat_output, history_state])
|
| 64 |
|
|
|
|
| 65 |
if __name__ == "__main__":
|
| 66 |
demo.launch()
|