Saya baru baca tulisan di AItidbids tentang state-of-the art di bidang generative AI. State-of-the-art itu bisa diartikan sebagai teknologi terkini, atau yang paling unggul. Jadi Aitidbits ini membuat daftar teknologi terkini di bidang generative AI tahun 2023, kayak semacam kaleidoskop.
Model bahasa (language model)
Model open source:
Yi: https://github.com/01-ai/Yi
Mistral: https://mistral.ai
Llama2: https://ai.meta.com/llama/
Code lama: https://ai.meta.com/blog/code-llama-large-language-model-coding/
Vicuna: https://lmsys.org/blog/2023-03-30-vicuna/
Benchmark model LLM:
Huggingface:
https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard
Hendrycks dkk
https://arxiv.org/abs/2009.03300
Truthful QA:
https://arxiv.org/abs/2109.07958
Model Berbayar:
GPT4 Turbo
Gemini Ultra
https://deepmind.google/technologies/gemini/#introduction
Claude 2.1
https://www.anthropic.com/index/claude-2-1
Multimodal
Open Source:
CogVLM
https://github.com/THUDM/CogVLM
Adept Fuyu
https://www.adept.ai/blog/fuyu-8b
ImigeBind
https://ai.meta.com/blog/imagebind-six-modalities-binding-ai/
AnyMal
https://arxiv.org/abs/2309.16058
Lava
data2vec:
https://ai.meta.com/blog/ai-self-supervised-learning-data2vec/
Commercial:
GPT & Gemini
Autonomous Agent:
Simulacra:
https://arxiv.org/abs/2304.03442
Agent Builder:
AutoGPT:
https://github.com/Significant-Gravitas/AutoGPT
BabyAGI:
https://github.com/yoheinakajima/babyagi
Agent-based app:
GPT Engineer:
https://github.com/gpt-engineer-org/gpt-engineer
Meta GPT
https://github.com/geekan/MetaGPT
Sota Model: CogAgent:
https://arxiv.org/abs/2312.08914
Berbayar:
Adept AI:
Embra:
Lindy:
Induced:
Hyperwrite:
https://www.hyperwriteai.com/personal-assistant
Sampe sini dulu nanti saya lanjutkan tentang State-of-the-art untuk image generation, video generation dll:
reference:
Tinggalkan Balasan