🤖 XWin LLM

Model by 🔗 Xwin-LM Team

@software{xwin-lm,
  title = {Xwin-LM},
  author = {Xwin-LM Team},
  url = {https://github.com/Xwin-LM/Xwin-LM},
  version = {pre-release},
  year = {2023},
  month = {9},
}

This model is based on Llama2 and was released under the same license. This model is just a LLM, meaning that this model only generates the most likely piece of text following a prompt. We use the 7B version of this model. Cavecats may include:

• generating text that is actually the prompt and then maybe answering the (modified) prompt
• not fully understanding the prompt
• writing very similar tweets in one generation
• making undesired comments and/or remarks

📥 Download

from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Xwin-LM/Xwin-LM-7B-V0.2")
model = AutoModelForCausalLM.from_pretrained("Xwin-LM/Xwin-LM-7B-V0.2")