chris_lee
chris_lee
这是使用chat.sh。 prompt type=0  这是prompt type=1 
# uninstall sdl2 brew uninstall sdl2 then install sdl2 from source code.... git clone https://github.com/libsdl-org/SDL.git -b SDL2 cd SDL mkdir build cd build ../configure make sudo make install
> > 我也遇到的相同的问题，在dify没有调用的时候。CPU 仍然居高不下。只能重启恢复 > > I also encounter the same problem. When Dify is not invoked, the CPU remains stubbornly high. Only a restart can restore it. > >...
> 我发现了两个可能与您的问题相关的类似讨论： > > 1. [如何把思考过程放到xinference中的DeepSeek R1部署中的思考中](https://github.com/langgenius/dify/discussions/13444)，目前尚未解决[[1]](https://github.com/langgenius/dify/discussions/13444)。 > 2. [使用DeepSeek的API，无法展示DeepSeek的思维过程](https://github.com/langgenius/dify/discussions/13125)，这也是一个未解决的问题[[2]](https://github.com/langgenius/dify/discussions/13125)。 > > 这些讨论解决了类似的担忧，即 DeepSeek-r1 模型输出中缺少“思考”内容。通过关注这些主题，您可能会发现更多见解或潜在解决方案。 > > _[要继续与Dosu](https://dosu.dev)交谈，请提及[@dosu](https://github.com/dosu)。_ > > 帮助 Dosu 学习！提供反馈：[很好的回应](https://app.dosu.dev/response-feedback/4ba4206e-da0a-4393-977b-18e9fba6ff3b?feedback_type=great_response)|[不相关的答案](https://app.dosu.dev/response-feedback/4ba4206e-da0a-4393-977b-18e9fba6ff3b?feedback_type=irrelevant_answer)|[不正确的来源](https://app.dosu.dev/response-feedback/4ba4206e-da0a-4393-977b-18e9fba6ff3b?feedback_type=incorrect_sources)|[太冗长](https://app.dosu.dev/response-feedback/4ba4206e-da0a-4393-977b-18e9fba6ff3b?feedback_type=too_verbose)|[幻觉](https://app.dosu.dev/response-feedback/4ba4206e-da0a-4393-977b-18e9fba6ff3b?feedback_type=hallucination)|[其他](https://app.dosu.dev/response-feedback/4ba4206e-da0a-4393-977b-18e9fba6ff3b?feedback_type=other)|[错误报告](https://app.dosu.dev/response-feedback/4ba4206e-da0a-4393-977b-18e9fba6ff3b?feedback_type=bug_report) But the latest version of dify official...
> 你需要检查结果的原始上下文，看看是否有think标签。  novita's supplier deepseek-r1 will output the Thinking tag。 but Openrouter's deepseek-r1 no Thinking content 。
> > Which model provider do you use? > > It seems duplicated with [#18866](https://github.com/langgenius/dify/issues/18866)   { "code": "invalid_param", "message": "1 validation error for File\n Value error, Invalid file...
> [@lichao4Java](https://github.com/lichao4Java) I am using the same model `gpt-4-1.nano`, it's working as expected. I used AWS S3 storage, does it have anything to do with this
Using APIs is normal。 such as: --data-raw '{ "inputs": {}, "query": "图片里有什么", "response_mode": "streaming", "conversation_id": "", "user": "abc-123", "files": [ { "type": "image", "transfer_method": "remote_url", "url": "https://www.baidu.com/img/PCtm_d9c8750bed0b3c7d089fa7d55720d6cf.png" } ] }'...
My previous version ran without any issues, including updates from previous versions. Why does this update require so many extra operations? I found that this issue is related to this...
``` File "/app/api/app.py", line 155, in initialize_extensions ext_storage.init_app(app) File "/app/api/extensions/ext_storage.py", line 65, in init_app storage.init_app(app) File "/app/api/extensions/ext_storage.py", line 22, in init_app self.storage_runner = S3Storage(app=app) File "/app/api/extensions/storage/s3_storage.py", line 33, in __init__...