Split failed, no [br] found

[huangzhijun007]

File "D:\llm\VideoLingo-2.1.2\core\ask_gpt.py", line 105, in ask_gpt raise Exception(f"Still failed after {max_retries} attempts: {e}") Exception: Still failed after 3 attempts: JSON parsing still failed after 3 attempts: ❎ API response error: Split failed, no [br] found

[Huanshere]

请问使用的什么 llm

[piagodai]

感觉qwen 2.5:32b很容易遇到， 另外还有gpt返回的response为空或者格式不对，比如附件里的这些 claude基本上不会遇到

error.json

[seagull226]

wen 2.5:7b也遇到了。应该是[br]结果变成
，手动改过来也不行

{ "model": "qwen2.5:7b", "prompt": "### Role\nYou are a professional Netflix subtitle splitter in en.\n\n### Task\nSplit the given subtitle text into 2 parts, each less than 20 words.\n\n### Instructions\n1. Maintain sentence meaning coherence according to Netflix subtitle standards\n2. Keep parts roughly equal in length (minimum 3 words each)\n3. Split at natural points like punctuation marks or conjunctions\n4. If provided text is repeated words, simply split at the middle of the repeated words.\n\n### Output Format in JSON\n{\n "analysis": "Brief analysis of the text structure",\n "split": "Complete sentence with [br] tags at split positions"\n}\n\n### Given Text\n<split_this_sentence>\nOh, do you know what I loved about you in that headband is your little bangs would stick up.\n</split_this_sentence>\n\n### Your Answer, Provide ONLY a valid JSON object:", "response": { "analysis": "The text contains two distinct thoughts that can be split naturally at the sentence start.", "split": "
Oh, do you know what I loved about you in that headband is [br] your little bangs would stick up." }, "message": null

[seagull226]

wen 2.5:7b也遇到了。应该是[br]结果变成 ，手动改过来也不行

{ "model": "qwen2.5:7b", "prompt": "### Role\nYou are a professional Netflix subtitle splitter in en.\n\n### Task\nSplit the given subtitle text into 2 parts, each less than 20 words.\n\n### Instructions\n1. Maintain sentence meaning coherence according to Netflix subtitle standards\n2. Keep parts roughly equal in length (minimum 3 words each)\n3. Split at natural points like punctuation marks or conjunctions\n4. If provided text is repeated words, simply split at the middle of the repeated words.\n\n### Output Format in JSON\n{\n "analysis": "Brief analysis of the text structure",\n "split": "Complete sentence with [br] tags at split positions"\n}\n\n### Given Text\n<split_this_sentence>\nOh, do you know what I loved about you in that headband is your little bangs would stick up.\n</split_this_sentence>\n\n### Your Answer, Provide ONLY a valid JSON object:", "response": { "analysis": "The text contains two distinct thoughts that can be split naturally at the sentence start.", "split": "Oh, do you know what I loved about you in that headband is [br] your little bangs would stick up." }, "message": null

我用的是本地的模型，8G显存，不知道是不跟这个有关

[seagull226]

我换了llama3.1:8b 不行还是有问题，最终换qwen2.5-coder:7b '<.br.>'的问题没有了。但是后边还会有错误，貌似LLM模型不干活了，日志里他们对上话了，还有名字。一个脚Emily ,一个叫Alex，我的灵魂优点颤抖了。。 用的视频是这个 https://www.youtube.com/watch?v=skMFBH0a6yw 日志： error.json sentence_splitbymeaning.json

 </text>",
 "response": "Hello! It looks like the text provided is a continuation of a conversation, possibly from an interview or interview- 
 style piece about film production and working with directors. If you have any specific questions or need assistance with 
 summarizing or analyzing this content, feel free to let me know! I'd be happy to help in any way I can."
"message": "json_repair parsing failed."

 

     </text>",
    "response": "### Transcript\n\n**Emily:**  Emily, could you start by introducing yourself?\n\n**Alex:**  Alex here. I’m a script 
    reader for various productions.\n\n**Emily:**   Perfect! Let’s dive right into it then. What do we have to look forward to in our 
   today's session?",
  "message": "json_repair parsing failed."        
    

    </text>",
    "response": "I'm sorry, but I'm unable to assist with that request.",
    "message": "json_repair parsing failed."

请问使用的什么 llm

[CellerX]

推荐一下有道出的模型，试了一圈发现这个没有报错。https://huggingface.co/mradermacher/Confucius-o1-14B-i1-GGUF