Beta
Explore
Marketplace
Neural Labs
Chat
Wallet
Docs
banban-vision-8b-v1.2.1-zh-DPO – AI Model by asadfgglie | AlphaNeural AI
You can deploy this model and start earning money today!
asadfgglie
/
banban-vision-8b-v1.2.1-zh-DPO
like
0
transformers
safetensors
llava
image-to-text
llama-factory
text-generation
conversational
asadfgglie/BanBan_2024-10-17-DPO
asadfgglie/banban-vision-8b-v1.2-zh
finetune
endpoints_compatible
us
Views
No views yet
Model card
Files and Versions
Community
API
Deploy
asadfgglie/banban-vision-8b-v1.2.1-zh-DPO
將asadfgglie/banban-vision-8b-v1.2-zh對asadfgglie/BanBan_2024-10-17-DPO再次DPO訓練後的版本
目前已知問題是中文能力會直接消失
英文能力反而有靠向這個資料集想要的偏好方向
推測是epoch=4導致的訓練過擬合
我原本想說eval loss與train loss都穩步下降的說
結果我等了10小時卻給我這個
哭阿
