Qinyuan Cheng

[email protected]

Fudan University&Sun Yat-Sen University Shanghai, China 萧瑟

Results 16 comments of


                                            Qinyuan Cheng

problem in prepro.py

> do you solve it ?can you tell me, thanks! Sry for answering too late, i have fixed vggnet.py to load the vggnet's parameters. You should fix the function build_params()...

problem in prepro.py

> thanks for your answering，I did this yesterday. but I have the following error > InvalidArgumentError: Shape must be rank 1 but is rank 2 for 'BiasAdd_7' (op: 'BiasAdd') with...

报错咋解决呀

请见 #161 #33

启动报错了 please pass along an `offload_folder`.

应该是显存不够的问题，请见#161

设置两张卡，只使用了一张卡，导致显存溢出报错

并行会先加载一张卡再加载另一张，是不是第一张卡剩余显存不够了，加载第一张卡的时候就OOM了

设置两张卡，只使用了一张卡，导致显存溢出报错

moss_gui_demo.py已经更新掉了，请重新pull一下最新的代码~

使用fnlp/moss-moon-003-sft-int4和fnlp/moss-moon-003-sft-int8模型报错

可以参考Readme中加载量化模型的步骤，并使用Huggingface Hub上最新的量化模型文件~

按照步骤安装完成运行moss_cli_demo.py报错

显存不够时需要在load_checkpoint_and_dispatch函数中指定一下offload_folder参数。

NameError: name 'autotune' is not defined

也可以尝试清一下本地的Huggingface的缓存，重新从Huggingface Hub上下载一下，模型文件有所更新，最新的模型文件直接按照Readme里的步骤跑量化模型应该ok的

求出个量化版的推理示例

插件版的推理示例还在整理中~

1
2
›