Upload 3 files

README.md

@@ -1,12 +1,14 @@
 ---
-title: Test 03
-emoji: 🐨
-colorFrom: purple
-colorTo: indigo
+title: TEST 01
+emoji: 🧠
+colorFrom: blue
+colorTo: red
 sdk: gradio
-sdk_version: 5.25.0
+sdk_version: 3.50.2
 app_file: app.py
 pinned: false
 ---
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+⚠️ 本模型為 gated model，請至 [Hugging Face token 設定頁](https://huggingface.co/settings/tokens) 創建 token，並在本 Space 的 Settings → Secrets 中添加 `HF_TOKEN`。
+
+---

app.py

@@ -0,0 +1,43 @@
+import gradio as gr
+import torch
+from PIL import Image
+from transformers import AutoProcessor, AutoModelForVision2Seq
+
+# 模型与处理器
+model_id = "Qwen/Qwen1.5-VL-Chat"
+processor = AutoProcessor.from_pretrained(model_id, trust_remote_code=True)
+model = AutoModelForVision2Seq.from_pretrained(
+    model_id,
+    device_map="auto",  # 自动分配 GPU 或 CPU
+    trust_remote_code=True,
+    torch_dtype=torch.bfloat16 if torch.cuda.is_available() else torch.float32,
+    **{"disable_exllama": True}  # 防止加载失败
+).eval()
+
+# 推理函数
+def chat(image, question):
+    if image is None or question.strip() == "":
+        return "請上傳圖片並輸入問題。"
+    inputs = processor(text=question, images=image, return_tensors="pt").to(model.device)
+    outputs = model.generate(**inputs, max_new_tokens=512)
+    answer = processor.batch_decode(outputs, skip_special_tokens=True)[0]
+    return answer.strip()
+
+# 界面設計
+with gr.Blocks(title="Qwen1.5-VL 圖文問答 Demo") as demo:
+    gr.Markdown("## 🧠 Qwen1.5-VL 圖文問答 Demo")
+    
+    with gr.Row():
+        with gr.Column():
+            image_input = gr.Image(type="pil", label="📷 請上傳圖片")
+            question_input = gr.Textbox(label="請輸入問題", placeholder="例如：這是什麼地方？")
+            submit_btn = gr.Button("Submit", variant="primary")
+            clear_btn = gr.Button("Clear")
+        with gr.Column():
+            answer_output = gr.Textbox(label="💬 答案", lines=8)
+
+    submit_btn.click(fn=chat, inputs=[image_input, question_input], outputs=answer_output)
+    clear_btn.click(lambda: ("", "", ""), outputs=[image_input, question_input, answer_output])
+
+# 啟動服務
+demo.launch(share=True)  # 如果你不想公開訪問可以改為 share=False

requirements.txt

@@ -0,0 +1,9 @@
+transformers>=4.36.2
+torch
+gradio
+einops
+tiktoken
+torchvision
+optimum
+auto-gptq
+transformers_stream_generator