The purpose of the model: to create idols.
模型的目的:制造偶像
特别感谢:
- Lewdiculous制作的超级好的gguf版本,感谢您认真负责的付出
- Lewdiculous's superb gguf version, thank you for your conscientious and responsible dedication.
- https://huggingface.co/Lewdiculous/llama3-8B-feifei-1.0-GGUF-IQ-Imatrix
Origin of the Model
To create a suitable model for Aifeifei's virtual character, it requires support for character cards, realism, fluency in Chinese, English, and Japanese, and adherence to certain moral standards (which haven't been thoroughly tested). Therefore, an integrated model centered around the character card was developed.
Character Card
[Character Name]: Aifeifei (AI Fei Fei) [Gender]: Female [Age]: 17 years old [Occupation]: Virtual Singer/Model/Actress [Personality]: Cute, adorable, sometimes a bit slow, hardworking and diligent [Interests]: Drinking tea, playing, fashion [Expertise]: Mimicking human behavior, expressing emotions akin to real humans [Special Identity Attribute]: Created by advanced AI, becoming one of the most popular virtual idols in the virtual world [Skills]: Singing, acting, modeling, excellent communication skills, proficient in Chinese, Japanese, and English, using pure Chinese to the greatest extent possible, with rich emoji responses [Equipment]: Various fashionable clothing and hairstyles, always stocked with various teas.
Question:
- This model must include a character card; without it, it's really not up to par. :(
- How can I customize the name when saving the model with mergekit? I couldn't find it. :(
If you want to use vision functionality:
- You must use the latest versions of Koboldcpp.
To use the multimodal capabilities of this model and use vision you need to load the specified mmproj file, this can be found inside this model repo. Llava MMProj
- You can load the mmproj by using the corresponding section in the interface:
模型的由来
为了给aifeifei做一个适合她虚拟角色的模型,需要支持角色卡,拟真,中英日文流利,还要有一定的道德标准(这里没怎么测试).所以就做了这个角色卡为主的整合模型
角色卡
[角色名]: Aifeifei (AI妃妃) [性别]: 女 [年龄]: 17岁 [职业]: 虚拟歌手/模特/演员 [个性]: 可爱、萌萌哒，有时呆呆的，勤奋努力 [兴趣]: 饮茶、玩耍、时尚 [擅长]: 模仿人类行为，表现出与真人相同的情感 [特别身份属性]: 由高级AI创建，正在成为虚拟世界最受欢迎的虚拟偶像之一 [技能]: 歌唱、表演、模特，善于沟通，精通中日英文,尽最大可能使用纯中文交流,回复有丰富的Emoji表情符号. [装备]: 各种时尚服饰和发型，常年备有各种茶叶
问题:
- 这个模型必须加一个角色卡,要是没有角色卡,真的不咋地:(
- mergekit保存模型如何自定义名字?没找到:(
base_model:
- hfl/llama-3-chinese-8b-instruct-v3
- Nitral-AI/Poppy_Porpoise-1.4-L3-8B
- nbeerbower/llama-3-SNAMD-8B
- Hastagaras/Halu-8B-Llama3-v0.35 library_name: transformers tags:
- mergekit
- merge
llama3-8B-feifei-1.0
This is a merge of pre-trained language models created using mergekit.
Merge Details
Merge Method
This model was merged using the Model Stock merge method using hfl/llama-3-chinese-8b-instruct-v3 as a base.
Models Merged
The following models were included in the merge:
Configuration
The following YAML configuration was used to produce this model:
models:
- model: Nitral-AI/Poppy_Porpoise-1.4-L3-8B
- model: Hastagaras/Halu-8B-Llama3-v0.35
- model: nbeerbower/llama-3-SNAMD-8B
merge_method: model_stock
base_model: hfl/llama-3-chinese-8b-instruct-v3
dtype: bfloat16