#103 · Primary category： Foundation Models
Orion
Orion-14B is a multilingual foundation model family with base, chat, long-context, quantized, RAG, and agent-tuned variants, excelling in Chinese, English, Japanese, and Korean.
Project last updated：06/03/24
GitHub Stars
811
Forks
59
Contributors
10
License
Apache-2.0
Why we included this project
Orion-14B is a multilingual model family trained from scratch on a 2.5T-token corpus covering Chinese, English, Japanese, and Korean, with especially strong results on Japanese and Korean benchmarks. The repo carries more than the base 14B model: a chat version, a long-context variant, a quantized build for lighter deployments, and fine-tuned models aimed at RAG and agent workflows. Teams evaluating open-weight options for these languages can use the family to compare how one set of weights behaves across different serving setups rather than hopping between unrelated checkpoints. The README also walks through inference with vLLM and llama.cpp, so you can see how the models would fit an existing stack before committing.
Articles for this project
No articles for this project yet.
To suggest a topic or contribute an article, contact us.
Related projects in this category
transformers
🤗 Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.
CLIP
CLIP (Contrastive Language-Image Pretraining), Predict the most relevant text snippet given an image
MiniCPM-V
A Pocket-Sized MLLM for Ultra-Efficient Image and Video Understanding on Your Phone
generative-models
Generative Models by Stability AI
unilm
Large-scale Self-supervised Pre-training Across Tasks, Languages, and Modalities