Delius

Results 7 comments of Delius

> > 能否提供下 tigerbot-7b-sft-4bit-128g 这个? 感谢 > > 两个7b的模型都给你了。 链接: https://pan.baidu.com/s/1WRqBLdmMZ_csagAwfkMsoQ?pwd=mth9 提取码: mth9 请问百度网盘有13b的吗?

> > 请问百度网盘有13b的吗? > > 链接: https://pan.baidu.com/s/1XhUrTDDcss3B321GJW_V7g?pwd=e9ny 提取码: e9ny --来自百度网盘超级会员v100的分享 感谢!

> 你好,使用如下配置和命令,是可以在4张3090上训练LLaMA-13B的。需要注意的是,该`batch_size`开得过大,并不能优化出WIC数据集上的最优结果。 > > ``` > # model > model_name_or_path: '/remote-home/share/llama_hf/13B' > # data > dataset_name: 'wic' > refresh: false > data_tag: 'base' > train_on_inputs: false > data_max_length: 1024 >...

> 可以参考https://www.deepspeed.ai/getting-started/#resource-configuration-multi-node 3张3090训练13B报OOM👇 ![f6e5cf36d76a53c8406474379c19ad6](https://github.com/OpenLMLab/LOMO/assets/88967316/0a276657-b23e-4e7b-b2a9-2dab50c60c85) ![21824581f1586f4099ee3cce12ca852](https://github.com/OpenLMLab/LOMO/assets/88967316/35f6ebee-35fc-4fdd-a687-317a9681b14b) 参数配置如下: args_lomo.yaml: ![5d2cf71a8467d2d7e6077dff8f7089a](https://github.com/OpenLMLab/LOMO/assets/88967316/53e1b7fe-8854-4ab4-aa23-87dae7b9e0da) ds_config.json: ![af3b0630917ff13060762871a1a7a48](https://github.com/OpenLMLab/LOMO/assets/88967316/f2fe36a7-994d-47a6-930c-8362e55a1543) run.sh: ![2105cb61cd9688667660d8376e61a0f](https://github.com/OpenLMLab/LOMO/assets/88967316/6a498183-1571-4896-afae-f4ece5a6def8) 跑得是baichuan-13b。 对源码的修改我就添加了loss在0.46以下时保存在一个特殊的output directory: ![e7e48cfac59a152ed071b7aa50c7d9b](https://github.com/OpenLMLab/LOMO/assets/88967316/04761f0d-a39c-4242-aadb-2177488bfee9) 这咋弄呀

> anyone meet this question? torch.onnx.errors.UnsupportedOperatorError: Exporting the operator 'aten::bincount' to ONNX opset version 12 is not supported. Please feel free to request support or submit a pull request on...