webui支持图片输入（仅多模态模型可用）

requirements-server.txt

@@ -1,4 +1,5 @@
 fastapi
 pydantic
 openai
-shortuuid
+shortuuid
+unicorn

tools/fastllm_pytools/web_demo.py

@@ -30,6 +30,8 @@ def get_model():
 
 if "messages" not in st.session_state:
     st.session_state.messages = []
+if "images" not in st.session_state:
+    st.session_state.images = []
 
 system_prompt = st.sidebar.text_input("system_prompt", "")
 max_new_tokens = st.sidebar.slider("max_new_tokens", 0, 8192, 512, step = 1)
@@ -43,12 +45,35 @@ def get_model():
     st.session_state.messages = []
     st.rerun()
 
+# 添加文件输入部件
+if (uploaded_file := st.file_uploader("上传图片", type=["jpg", "jpeg", "png"])) is not None:
+    from PIL import Image
+    import io
+    import base64
+
+    image = Image.open(uploaded_file).convert('RGB')
+    st.session_state.images = [image]
+
+    buffered = io.BytesIO()
+    image.save(buffered, format="PNG")
+    img_str = base64.b64encode(buffered.getvalue()).decode()
+
+    # 创建一个包含图片的 HTML 元素，并设置样式
+    image_html = f"""
+    <div style="display: flex; justify-content: flex-end;">
+        <img src="data:image/png;base64,{img_str}" style="max-width: 300px; max-height: 300px;">
+    </div>
+    """
+
+    # 显示图片
+    st.markdown(image_html, unsafe_allow_html=True)
+
 for i, (prompt, response) in enumerate(st.session_state.messages):
     with st.chat_message("user"):
         st.markdown(prompt)
     with st.chat_message("assistant"):
         st.markdown(response)
-
+    
 if prompt := st.chat_input("请开始对话"):
     model = get_model()
     with st.chat_message("user"):
@@ -65,14 +90,24 @@ def get_model():
             messages.append({"role": "assistant", "content": his[1]})
         messages.append({"role": "user", "content": prompt})
 
-        for chunk in model.stream_response(messages,
+        if (len(st.session_state.images) > 0):
+            handle = model.launch_stream_response(messages,
+                        max_length = max_new_tokens, do_sample = True,
+                        top_p = top_p, top_k = top_k, temperature = temperature,
+                        repeat_penalty = repeat_penalty, one_by_one = True, images = st.session_state.images)
+            for chunk in model.stream_response_handle(handle):
+                full_response += chunk
+                message_placeholder.markdown(full_response + "▌")
+            message_placeholder.markdown(full_response)
+        else:
+            for chunk in model.stream_response(messages,
                                            max_length = max_new_tokens,
                                            top_k = top_k,
                                            top_p = top_p,
                                            temperature = temperature,
                                            repeat_penalty = repeat_penalty,
                                            one_by_one = True):
-            full_response += chunk
-            message_placeholder.markdown(full_response + "▌")
-        message_placeholder.markdown(full_response)
+                full_response += chunk
+                message_placeholder.markdown(full_response + "▌")
+            message_placeholder.markdown(full_response)
     st.session_state.messages.append((prompt, full_response))