Update Inference.py

Inference.py

@@ -6,6 +6,7 @@ from tensorflow.keras import mixed_precision
 import sentencepiece as spm
 import os, json
 import requests
+import gradio as gr
 
 print('1')
 
@@ -57,7 +58,7 @@ if not os.path.exists(MODEL_PATH):
 
 if not os.path.exists(TOKENIZER_PATH):
     download_file(
-        "https://huggingface.co/datasets/Yuchan5386/TinyInst/resolve/main/ko_unigram.model?download=true",
+        "https://huggingface.co/Yuchan5386/Respiso/resolve/main/bpe.model?download=true",
         TOKENIZER_PATH
     )
 
@@ -225,5 +226,23 @@ def generate_text_topp(model, prompt, max_len=512, max_gen=512, p=0.9, temperatu
         generated.append(int(next_token_id))
     return ids_to_text(generated)
 
-print("\n\n===== 생성 결과 =====")  
-print(generate_text_topp(model, "지난 2년 동안", p=0.8))
+def gr_generate(prompt, max_len=512, max_gen=512, p=0.8, temperature=0.8):
+    return generate_text_topp(model, prompt, max_len=max_len, p=p, temperature=temperature)
+
+# Gradio 인터페이스 정의
+iface = gr.Interface(
+    fn=gr_generate,
+    inputs=[
+        gr.Textbox(label="Prompt 입력", placeholder="여기에 문장 입력...", lines=2),
+        gr.Slider(20, 512, value=150, step=1, label="Max length"),
+        gr.Slider(0.1, 1.0, value=0.8, step=0.05, label="Top-p"),
+        gr.Slider(0.1, 2.0, value=0.8, step=0.05, label="Temperature")
+    ],
+    outputs=[
+        gr.Textbox(label="생성 결과", lines=10)
+    ],
+    title="Cuma LM 텍스트 생성",
+    description="간단한 Gradio UI로 Cuma 모델 텍스트 생성 테스트"
+)
+
+iface.launch()