SkyworkAI / Skywork

Skywork series models are pre-trained on 3.2TB of high-quality multilingual (mainly Chinese and English) and code data. We have open-sourced the model, training data, evaluation data, evaluation methods, etc. 天工系列模型在3.2TB高质量多语言和代码数据上进行预训练。我们开源了模型参数,训练数据,评估数据,评估方法。
Other
1.21k stars 111 forks source link

fix domain name typo w.r.t ChineseDomainModelingEval #56

Closed ccclyu closed 8 months ago

ccclyu commented 9 months ago

Follow the domain names from https://huggingface.co/datasets/Skywork/ChineseDomainModelingEval for loss evaluation

ccclyu commented 8 months ago

@zhao1iang @TianwenWei what happened to this pull request? Just ignore and revise by yourself. Does your project welcome contributions from community?

zhao1iang commented 8 months ago

@ccclyu Apologies for the previous oversight where we combined multiple changes into a single pull request during the merge. We appreciate your feedback, and the community is encouraged to keep offering suggestions.