JittorLLMs

JittorLLMs copied to clipboard

chatglm第三个checkpoint超时

2

兄弟们，还有救吗  每次都是到第三个这里就会超时,同时有cuda错误，会不会是电脑性能的问题？（16g内存，4g显存） 恳请各位大佬指教

fi5ee

`python web_demo.py llama2` This path may cause jittor found the wrong libs, please unset LD_LIBRARY_PATH and remove cuda lib path in Path. ... RuntimeError: [f 0120 11:01:54.242000 48 compile_extern.py:247] CUDA...

showkeyjar

使用部署命令出现“'pangualpha' 不是内部或外部命令，也不是可运行的程序”错误

2

(JittorLLMs) C:\o\JittorLLMs>python cli_demo.py [chatglm|pangualpha|llama|chatrwkv] 'pangualpha' 不是内部或外部命令，也不是可运行的程序 或批处理文件。 (JittorLLMs) C:\o\JittorLLMs>python cli_demo.py [chatglm|pangualpha|llama|chatrwkv]python cli_demo.py [chatglm|pangualpha|llama|chatrwkv]] 'pangualpha' 不是内部或外部命令，也不是可运行的程序 或批处理文件

wsmdn

fix bug in LLaMAModel.chat() in __init__.py This line of code is missing an indentation, which results in "UnboundLocalError: local variable'output text' referenced before assignment"

yjcexp

运行时显示Unable to alloc cuda device memory, use unify memory instead. This may cause low performance

3

ubantu下 运行chatglm 加载不到显存中

HJT9328

1. 没有GPU，纯CPU 2. win11 WSL2 ubuntu 3. Python 3.8.16 4. pip -r requirements.txt 正常 5. python api.py chatglm 报错如下： /home/tom/miniconda3/lib/python3.8/site-packages/jittor/src/misc/cstr.h(19): error: no instance of overloaded function "std::unique_ptr::reset [with _Tp=char, _Dp=std::default_delete]"...

atomsi

post_data = json.dumps({'prompt': 'Hello, solve 5x=13'}) >>> print(json.loads(requests.post("http://0.0.0.0:8000", post_data).text)['response']) Traceback (most recent call last): File "", line 1, in File "D:\software\python3.8\lib\json\__init__.py", line 357, in loads return _default_decoder.decode(s) File "D:\software\python3.8\lib\json\decoder.py", line...

dididiskq

报错： in check_cuda assert ret == 0 AssertionError

1

PS D:\samples\llm\JittorLLMs> python cli_demo.py chatrwkv [i 0814 09:11:50.124000 64 compiler.py:956] Jittor(1.3.8.5) src: c:\users\aaa\appdata\local\programs\python\python310\lib\site-packages\jittor [i 0814 09:11:50.153000 64 compiler.py:957] cl at C:\Users\AAA\.cache\jittor\msvc\VC\_\_\_\_\_\bin\cl.exe(19.29.30133) [i 0814 09:11:50.154000 64 compiler.py:958] cache_path: C:\Users\AAA\.cache\jittor\jt1.3.8\cl\py3.10.1\Windows-10-10.x86\IntelRCoreTMi7x93\default [i 0814...

linfx

在最后一步python cli_demo.py报错求助

2

Building wheels for collected packages: jtorch, torch, torchvision Building wheel for jtorch (setup.py) ... done Created wheel for jtorch: filename=jtorch-0.1.7-py3-none-any.whl size=52207 sha256=0dcf8be047ce00a1b4878ec767167e0c54f279606b4350c5ec4c00c820cdb647 Stored in directory: c:\users\liuti\appdata\local\pip\cache\wheels\86\6e\e8\ee25ca9857bb3b1d6a7fd285dc3b88492e65df5e76ffc4507e Building wheel for torch...

liutianye

出现load parameter rope.freqs failed ...的报错，请问怎么处理呢

yjcexp