Update app.py

app.py

@@ -1,10 +1,19 @@
 import streamlit as st
-from transformers import pipeline
-from gtts import gTTS
-import io
-import os
-import langdetect
-import tempfile  # 添加这一行导入 tempfile 模块
+import importlib.util
+try:
+    # 检查 accelerate 库是否安装
+    spec = importlib.util.find_spec("accelerate")
+    if spec is None:
+        st.error("缺少 'accelerate' 库，请安装该库以加载 FP8 量化模型。可以使用 'pip install accelerate' 进行安装。")
+        st.stop()
+    from transformers import pipeline
+    from gtts import gTTS
+    import io
+    import tempfile
+    import os
+except ImportError as e:
+    st.error(f"导入库时出错: {e}")
+    st.stop()
 
 # function part
 # img2text
@@ -14,42 +23,30 @@ def img2text(url):
         text = image_to_text_model(url)[0]["generated_text"]
         return text
     except Exception as e:
-        st.error(f"图像转文本出错: {e}")
+        st.error(f"图像描述生成出错: {e}")
         return None
 
-# text2story
+#  text2story 
 def text2story(text):
     try:
-        story_generator = pipeline("text-generation", model="Qwen/QwQ-32B")
-        story = story_generator(text, max_length=200, num_return_sequences=1)[0]["generated_text"]
+        
+        story_generator = pipeline("text-generation", model="perplexity-ai/r1-1776", trust_remote_code=True)
+        story = story_generator(text, max_length=200, num_return_sequences=1)[0]['generated_text']
         return story
     except Exception as e:
-        st.error(f"文本生成故事出错: {e}")
+        st.error(f"故事生成出错: {e}")
         return None
 
 # text2audio
 def text2audio(story_text):
     try:
-        # 检测故事的语言
-        detected_lang = langdetect.detect(story_text)
-        # 根据检测到的语言创建 gTTS 对象
-        tts = gTTS(text=story_text, lang=detected_lang)
-        # 创建一个字节流对象用于存储音频数据
-        audio_file = io.BytesIO()
-        # 将音频数据写入字节流
-        tts.write_to_fp(audio_file)
-        # 将文件指针移到开头
-        audio_file.seek(0)
-        return audio_file
-    except langdetect.LangDetectException:
-        st.error("无法检测故事的语言，默认使用英语进行语音合成。")
         tts = gTTS(text=story_text, lang='en')
         audio_file = io.BytesIO()
         tts.write_to_fp(audio_file)
         audio_file.seek(0)
         return audio_file
     except Exception as e:
-        st.error(f"文本转音频出错: {e}")
+        st.error(f"文本转语音出错: {e}")
         return None
 
 st.set_page_config(page_title="Your Image to Audio Story",
@@ -65,21 +62,21 @@ if uploaded_file is not None:
         temp_file_path = temp_file.name
 
     st.image(uploaded_file, caption="Uploaded Image",
-             use_container_width=True)
+             use_container_width=True)  # 修改为 use_container_width
 
-    # Stage 1: Image to Text
+    #Stage 1: Image to Text
     st.text('Processing img2text...')
     scenario = img2text(temp_file_path)
     if scenario:
         st.write(scenario)
 
-        # Stage 2: Text to Story
+        #Stage 2: Text to Story
         st.text('Generating a story...')
         story = text2story(scenario)
         if story:
             st.write(story)
 
-            # Stage 3: Story to Audio data
+            #Stage 3: Story to Audio data
             st.text('Generating audio data...')
             audio_data = text2audio(story)
             if audio_data:
@@ -89,8 +86,5 @@ if uploaded_file is not None:
                              format="audio/mpeg",
                              start_time=0)
 
-    # 删除临时文件并进行异常处理
-    try:
-        os.remove(temp_file_path)
-    except Exception as e:
-        st.error(f"删除临时文件出错: {e}")
+    # 删除临时文件
+    os.remove(temp_file_path)