GalSang17

Results 11 comments of GalSang17

Hi, Now I have a new issue. When I used the file **main_test_table1.py** to test my own datasets. When the programming run a little while, it can catch exception like...

换到这个低版本的还是不行

> 支持Alpaca等指令数据集的SFT和RLHF流程:https://github.com/hiyouga/LLaMA-Efficient-Tuning > > 运行以下指令即可实现 Alpaca 数据集指令微调(instruction-tuning): > > ```shell > CUDA_VISIBLE_DEVICES=0 python src/train_sft.py \ > --model_name_or_path baichuan-7B模型文件夹路径 \ > --do_train \ > --dataset alpaca_gpt4_zh \ > --finetuning_type lora \...

> @GalSang17 项目自带了,点进data文件夹就可以看示例格式。 谢谢!

![image](https://github.com/baichuan-inc/baichuan-7B/assets/49866330/f837f8bb-4520-4aff-b7b7-5e236e40cddd)

> Hi @zhudu , Thanks for pointing it out. We will rectify it . You can reproduce the accuracy by running python val.py --data coco.yaml --img 640 --conf 0.001 --iou...

I am also using this implementation, is it impossible to call distributed training?

> 那一版固定了15个类别,需要在网络那里修改自定义类别数量。我修好好了但还没时间上传 > […](#) > ---Original--- From: "Qing ***@***.***> Date: Tue, Aug 20, 2024 14:27 PM To: ***@***.***>; Cc: ***@***.***>;"State ***@***.***>; Subject: Re: [wang-xinyu/tensorrtx] Dev obb (PR #1539) 你好 尝试你的...

@yzqxy 我就是用的v8n的yaml文件呢,稀疏训练的配置是这样 ![image](https://github.com/yzqxy/Yolov8_obb_Prune_Track/assets/49866330/1ba4cd00-73b6-4373-9c5b-83639e14e5a5) 然后test-p.yaml里面是这样子 ![image](https://github.com/yzqxy/Yolov8_obb_Prune_Track/assets/49866330/7f06fbf6-b525-4111-981a-e97903bf211d) 你看方便加个微信吗,我有一些剪枝的问题想向你请教呢^_^,J-M116

> > @yzqxy 我就是用的v8n的yaml文件呢,稀疏训练的配置是这样 ![image](https://user-images.githubusercontent.com/49866330/267878374-1ba4cd00-73b6-4373-9c5b-83639e14e5a5.png) 然后test-p.yaml里面是这样子 ![image](https://user-images.githubusercontent.com/49866330/267880130-7f06fbf6-b525-4111-981a-e97903bf211d.png) 你看方便加个微信吗,我有一些剪枝的问题想向你请教呢^_^,J-M116 > > ![image](https://user-images.githubusercontent.com/51917814/267881516-454aa7b6-0338-48f3-8e30-13f3ab9206cd.png) 我这边n是可行的,你再确认一下,你训练的yaml是不是n,剪枝命令也得是n ![image](https://github.com/yzqxy/Yolov8_obb_Prune_Track/assets/49866330/fca5ce03-3515-4f17-9f46-bccab5855586) 再次确认了,就是v8n.yaml,配置截图和报错如下 ![image](https://github.com/yzqxy/Yolov8_obb_Prune_Track/assets/49866330/4d470f3c-7209-42a5-9d41-ae06995c7eeb) 不过我稀疏训练结束时验证图片的预测效果好像偏移不对,像这样 ![image](https://github.com/yzqxy/Yolov8_obb_Prune_Track/assets/49866330/f3f03da4-195b-4a2d-8eb6-dd1a3d48848a) 训练的命令都是按照你的博客来的,单卡训练的