Open hiyin opened 11 months ago
同学您好,请问上传pdf后一直在processing是为什么呢?
你看下你的PDF parser的后台,感觉像是PDF的处理那边有些问题,我这边是Okay的,会给出feedback!
sciencebeam_parser.service.server
还有可能是你的GPT API的连接问题,都有可能
https://9dc2633eeb8fefc5c0.gradio.live @WangzcBruce 你可以试试我的暂时的,72小时就失效了。
好的,谢谢您。这里是我执行 sciencebeam_parser.service.server 的显示
https://9dc2633eeb8fefc5c0.gradio.live @WangzcBruce 你可以试试我的暂时的,72小时就失效了。 谢谢您!不过访问不了了, 可能key次数过多了 Failed to generate review... Error: The OpenAI account associated with this API key has been deactivated. If you are the developer for this OpenAI app, please check your email for more information. If you are seeing this error while using another app or site, please reach out to them for more help.
你试试如果加入share=TRUE进入lauch(),用public接口会好些吗
sciencebeam_parser.service.server 好像跟我的运行差不多的,是返回response的时候的问题
你试试如果加入share=TRUE进入lauch(),用public接口会好些吗
发现key一直都没有被调用 用来public接口也是一直转圈
想问一下正常会出现上传文件一直在uploading状态吗,20MB的pdf过了10分钟还在uploading,请问文件大小有限制吗?
我找到问题所在了,代码所在的服务器需要能够翻墙。感谢作者的帮助
想问一下正常会出现上传文件一直在uploading状态吗,20MB的pdf过了10分钟还在uploading,请问文件大小有限制吗?
pdf压缩一下
你试试如果加入share=TRUE进入lauch(),用public接口会好些吗
Thanks the author's help! https://385a2867849723de72.gradio.live/ It works now!!!!
For reference:
demo.launch(server_name="0.0.0.0", server_port=7799, share=True)
export https_proxy=socks5://xxx.xxx.xxx.xxx:xxx
, set export no_proxy="localhost, 127.0.0.1, ::1"
Hi, I encountered this error with all the documents I uploaded. I was wondering if there's any solution for this? Thanks!
你好,请问在llm中输入python main.py后报如下的错误:
你好,请问在llm中输入python main.py后报如下的错误:
请问这种情况是什么原因导致的呢
I found this tool extremely helpful and fun to use! Thanks to the developers you are awesome! However I believe the repository needs some care afterwards to make the documentation more user-approachable and that users can run the tool smoothly. I am speaking from my own user experience.
Here is the tips how I setup mine and successfully running:
- Git clone the repo in a Linux machine! (No Mac) I used CenOS 7 in our school
- Make sure you use the GPT-4 API Key, it starts with sk-xxxxxxx... (for mainland users maybe try purchase it from some online e-commerce shop), because I used the GPT-3.5-turbo API key and the quote just exceeded with no reason whenever I run. Also don't forget to change GPT-3.5-turbo to GPT-4 so your lines in main.py should be:
def __init__(self, model_name="gpt-4"):
wrapper = GPT4Wrapper(model_name="gpt-4")
- There is no
but you only need 4 packages stated below to make the main.py work i.e. gradio pikepdf openai tiktoken (you can also create your own requirement.txt with the four lines above) - If you are use school server make sure you change the launch() in the main.py to include share=TRUE to be able to use a public address for accessing the tool's web interface
- Upload your PDF and get some cool AI feedback!
3 #2 #1
You can use the "gpt-3.5-turbo-16k" instead. GPT4 is too expensive.
Thanks. I made it work with gpt 4 turbo on Ubuntu:
https://github.com/harrywang/LLM-scientific-feedback
I have made the following changes to the code base:
openai==0.28
to avoid this: https://github.com/Weixin-Liang/LLM-scientific-feedback/issues/25demo.launch(server_name="0.0.0.0", server_port=7799, share='True')
to main.py
to get a public URLdemo.queue(concurrency_count=3)
to demo.queue()
to remove a deprecation errorkey.txt
and put your API key as the only line in the file, such as sk-xxx
System Used: Ubuntu 20.04.6 LTS
conda env create -f conda_environment.yml
conda activate ScienceBeam
python -m sciencebeam_parser.service.server --port=8080 # Make sure this is running in the background
Then,
conda create -n llm python=3.10
conda activate llm
pip install -r requirements.txt
python main.py
Go to something like https://5a1xxxx.gradio.live/ and upload a PDF paper:
@hiyin 您好这个代码就这么多吗?看您复现成功了。我怎么觉得少了很多代码
I found this tool extremely helpful and fun to use! Thanks to the developers you are awesome! However I believe the repository needs some care afterwards to make the documentation more user-approachable and that users can run the tool smoothly. I am speaking from my own user experience.
Here is the tips how I setup mine and successfully running:
def __init__(self, model_name="gpt-4"):
wrapper = GPT4Wrapper(model_name="gpt-4")
3
2
1