[Bug]: 精准翻译PDF文档(NOUGAT)功能异常。

[ipc-robot]

Installation Method | 安装方法与平台

Pip Install (I used latest requirements.txt)

Version | 版本

Latest | 最新版

OS | 操作系统

Windows

Describe the bug | 简述

平台traceback [Local Message] 插件调用出错:

Traceback (most recent call last): File ".\toolbox.py", line 203, in decorated yield from f(main_input, llm_kwargs, plugin_kwargs, chatbot_with_cookie, history, *args, **kwargs) File ".\crazy_functions\批量翻译PDF文档_NOUGAT.py", line 93, in 批量翻译PDF文档 yield from 解析PDF_基于NOUGAT(file_manifest, project_folder, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt) File ".\crazy_functions\批量翻译PDF文档_NOUGAT.py", line 111, in 解析PDF_基于NOUGAT fpp = yield from nougat_handle.NOUGAT_parse_pdf(fp, chatbot, history) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File ".\crazy_functions\crazy_utils.py", line 599, in NOUGAT_parse_pdf raise RuntimeError("Nougat解析论文失败。") RuntimeError: Nougat解析论文失败。

Screen Shot | 有帮助的截图

Terminal Traceback & Material to Help Reproduce Bugs | 终端traceback（如有） + 帮助我们复现的测试材料样本（如有）

终端traceback D:\anaconda3\envs\sbgithub\Lib\site-packages\torch\functional.py:507: UserWarning: torch.meshgrid: in an upcoming release, it will be required to pass the indexing argument. (Triggered internally at C:\cb\pytorch_1000000000000\work\aten\src\ATen\native\TensorShape.cpp:3550.) return _VF.meshgrid(tensors, **kwargs) # type: ignore[attr-defined] 0%| | 0/3 [00:00<?, ?it/s] Traceback (most recent call last): File "<frozen runpy>", line 198, in _run_module_as_main File "<frozen runpy>", line 88, in _run_code File "D:\anaconda3\envs\sbgithub\Scripts\nougat.exe\__main__.py", line 7, in <module> File "D:\anaconda3\envs\sbgithub\Lib\site-packages\predict.py", line 167, in main model_output = model.inference( ^^^^^^^^^^^^^^^^ File "D:\anaconda3\envs\sbgithub\Lib\site-packages\nougat\model.py", line 592, in inference decoder_output = self.decoder.model.generate( ^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "D:\anaconda3\envs\sbgithub\Lib\site-packages\torch\utils\_contextlib.py", line 115, in decorate_context return func(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^ File "D:\anaconda3\envs\sbgithub\Lib\site-packages\transformers\generation\utils.py", line 1527, in generate result = self._greedy_search( ^^^^^^^^^^^^^^^^^^^^ File "D:\anaconda3\envs\sbgithub\Lib\site-packages\transformers\generation\utils.py", line 2408, in _greedy_search model_inputs = self.prepare_inputs_for_generation(input_ids, **model_kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ TypeError: BARTDecoder.prepare_inputs_for_inference() got an unexpected keyword argument 'cache_position' -> Cannot close object, library is destroyed. This may cause a memory leak! -> Cannot close object, library is destroyed. This may cause a memory leak! -> Cannot close object, library is destroyed. This may cause a memory leak! -> Cannot close object, library is destroyed. This may cause a memory leak! -> Cannot close object, library is destroyed. This may cause a memory leak! -> Cannot close object, library is destroyed. This may cause a memory leak!

这是我试图翻译的pdf文章： example.pdf

[SpectreLucifer]

transformers包版本不对，用4.39.0以上好像就会出这个问题

[SpectreLucifer]

我今天也是卡这了，花了几个钟头，服了

[ipc-robot]

我今天也是卡这了，花了几个钟头，服了

确实是这样的，我用4.39.0以下的包可以正常运行了！ 但这个NOUGAT翻译的结果和“精准翻译PDF论文”有啥区别。

[SpectreLucifer]

目前看来好像NOUGAT多一个latex公式

[hemingchun]

遇到同样的问题，尴尬。。。

[Neko-coffee]

这个问题具体怎么解决可以说一下嘛，遇到了同样个问题

[SpectreLucifer]

这个问题具体怎么解决可以说一下嘛，遇到了同样个问题

不太清楚现在修复了没有 如果没有，把transformers包换一下就行

pip install transformers==4.38.2

[Neko-coffee]

这个问题具体怎么解决可以说一下嘛，遇到了同样的问题

不太清楚现在修复了没有 如果没有，把变形金刚包换一下就行

pip install transformers==4.38.2

好的谢谢你