JackieWu

Results 12 comments of JackieWu

some questions in application

本工具集主要提供一些baseline模型，方便开发和对比模型结果，保存模型可以用于其他应用，模型导出和部署，请参考tensorflow相关的文档。

Instance key should be "query_id" when reader is CMRC

thank you for your question. You are right the key should by 'query_id', we will update it in later version.

启动之后，Human内输入，出现报错。如下

同样问题，求解答

160G内存，两张24G3090，800G硬盘的环境下，对GLM-10-chinese进行finetune

遇到同样问题

160G内存，两张24G3090，800G硬盘的环境下，对GLM-10-chinese进行finetune

@RileyShe 多卡只要修改ds_fintune_seq2seq 里面的配置就可以了，比如num_nodes,num_gpus 就可以了

小数据finetune large-chinese rouge 为0

应该不是这个原因，因为所有的pred过了判断重复的预测都是为空，所以rouge为0，我想在小数据上过拟合来着，调整学习率也不行

ddp traing multi gpu Expected all tensors to be on the same device, but found at least two devices

runing command as: ! CUDA_VISIBLE_DEVICES=0,1 python -m tevatron.driver.train \ --output_dir model_msmarco \ --model_name_or_path bert-base-uncased \ --save_steps 1000 \ --train_dir /kaggle/working/train_tevatron_100.json \ --fp16 \ --per_device_train_batch_size 2 \ --train_n_passages 8 \ --learning_rate...

missing pad_token error when using GPT2Chinese

Hi, @reciprocated ,trlx==0.6.0, tokenizer was loaded with AutoTokenizer

batch size 未定义

gpu环境

[Bug] Orpheus_tts espanish finetune ,cannot generate valid voice

Sorry for the late reply. this is the notebook i used, https://colab.research.google.com/drive/1sINXJCjZFPQDtUD9nYBdS3vRkKndLgsJ?usp=sharing The data i used can be downloaded from this link: https://drive.google.com/file/d/1rpMGSQLcdas9oM6xQ31rsvtp_KU7BxtF/view?usp=sharing I'm new to speech fine-tuning，would greatly appreciate...