Xwin LM 13B V0.2

🚀 Xwin-LM: Powerful, Stable, and Reproducible LLM Alignment

Xwin-LM is dedicated to developing and open-sourcing alignment technologies for large language models. These technologies include supervised fine-tuning (SFT), reward models (RM), reject sampling, and reinforcement learning from human feedback (RLHF). Our first release, based on the Llama2 base models, ranked TOP-1 on AlpacaEval, and it was the first to surpass GPT-4 on this benchmark. The project will be continuously updated.

🚀 Quick Start

To get started with Xwin-LM, you can follow the inference examples provided later in this README. Make sure to adhere to the conversation template for optimal results.

✨ Features

• Advanced Alignment Technologies: Xwin-LM incorporates various alignment techniques such as SFT, RM, reject sampling, and RLHF.
• Top Performance on Benchmarks: Our models have achieved top rankings on AlpacaEval, outperforming many well - known models including GPT - 4 in some cases.
• Multi - turn Conversation Support: The model supports multi - turn conversations following the prompt format established by Vicuna.

📦 Installation

No specific installation steps are provided in the original README. If you want to use the model for inference, you can refer to the HuggingFace or vLLM examples below.

💻 Usage Examples

Basic Usage

Conversation Template

To obtain desired results, please strictly follow the conversation templates when utilizing our model for inference. Our model adopts the prompt format established by Vicuna and is equipped to support multi - turn conversations.

A chat between a curious user and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the user's questions. USER: Hi! ASSISTANT: Hello.</s>USER: Who are you? ASSISTANT: I am Xwin-LM.</s>......

HuggingFace Example

from transformers import AutoTokenizer, AutoModelForCausalLM

model = AutoModelForCausalLM.from_pretrained("Xwin-LM/Xwin-LM-7B-V0.1")
tokenizer = AutoTokenizer.from_pretrained("Xwin-LM/Xwin-LM-7B-V0.1")
(
    prompt := "A chat between a curious user and an artificial intelligence assistant. "
            "The assistant gives helpful, detailed, and polite answers to the user's questions. "
            "USER: Hello, can you help me? "
            "ASSISTANT:"
)
inputs = tokenizer(prompt, return_tensors="pt")
samples = model.generate(**inputs, max_new_tokens=4096, temperature=0.7)
output = tokenizer.decode(samples[0][inputs["input_ids"].shape[1]:], skip_special_tokens=True)
print(output) 
# Of course! I'm here to help. Please feel free to ask your question or describe the issue you're having, and I'll do my best to assist you.

vLLM Example

Because Xwin-LM is based on Llama2, it also offers support for rapid inference using vLLM. Please refer to vLLM for detailed installation instructions.

from vllm import LLM, SamplingParams
(
    prompt := "A chat between a curious user and an artificial intelligence assistant. "
            "The assistant gives helpful, detailed, and polite answers to the user's questions. "
            "USER: Hello, can you help me? "
            "ASSISTANT:"
)
sampling_params = SamplingParams(temperature=0.7, max_tokens=4096)
llm = LLM(model="Xwin-LM/Xwin-LM-7B-V0.1")
outputs = llm.generate([prompt,], sampling_params)

for output in outputs:
    prompt = output.prompt
    generated_text = output.outputs[0].text
    print(generated_text)

📚 Documentation

News

• 💥 [Oct 12, 2023] Xwin-LM-7B-V0.2 and Xwin-LM-13B-V0.2 have been released, with improved comparison data and RL training (i.e., PPO). Their winrates v.s. GPT-4 have increased significantly, reaching 59.83% (7B model) and 70.36% (13B model) respectively. The 70B model will be released soon.
• 💥 [Sep, 2023] We released Xwin-LM-70B-V0.1, which has achieved a win - rate against Davinci-003 of 95.57% on AlpacaEval benchmark, ranking as TOP-1 on AlpacaEval. It was the FIRST model surpassing GPT-4 on AlpacaEval. Also note its winrate v.s. GPT-4 is 60.61.
• 🔍 [Sep, 2023] RLHF plays crucial role in the strong performance of Xwin-LM-V0.1 release!
• 💥 [Sep, 2023] We released Xwin-LM-13B-V0.1, which has achieved 91.76% win - rate on AlpacaEval, ranking as top-1 among all 13B models.
• 💥 [Sep, 2023] We released Xwin-LM-7B-V0.1, which has achieved 87.82% win - rate on AlpacaEval, ranking as top-1 among all 7B models.

Model Card

Benchmarks

Xwin-LM performance on AlpacaEval

The table below displays the performance of Xwin-LM on AlpacaEval, where evaluates its win - rate against Text-Davinci-003 across 805 questions. To provide a comprehensive evaluation, we present, for the first time, the win - rate against ChatGPT and GPT-4 as well. Our Xwin-LM model family establish a new state - of - the - art performance across all metrics. Notably, Xwin-LM-70B-V0.1 has eclipsed GPT-4 for the first time, achieving an impressive win - rate of 95.57% to Text-Davinci-003 and 60.61% to GPT-4.

Case study on AlpacaEval

Xwin-LM performance on NLP foundation tasks

The following table provides a comparison of Xwin-LMs with other LLMs on NLP foundation tasks in Open LLM Leaderboard.

📄 License

The Xwin-LM models are released under the Llama 2 License.

@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},
}

Thanks to Llama 2, FastChat, AlpacaFarm, and vLLM.