Ring-mini-2.0

🤗 Hugging Face   |   🤖 ModelScope

Introduction

We present a compact yet powerful reasoning model Ring-mini-2.0. It has 16B total parameters, with 1.4B parameters are activated per input token (non-embedding 789M). Although Ring-mini-2.0 is quite compact, it still reaches the top-tier level of sub-10B dense LLMs and even matches or surpasses much larger MoE models, through pre-training on 20T tokens of high-quality data and enhanced through long-cot supervised fine-tuning and multi-stage reinforcement learning.

Model Downloads

Model	#Total Params	#Activated Params	Context Length	Download
Ring-mini-2.0	16.8B	1.4B	128K	🤗 HuggingFace
Ring-lite-2507	16.8B	2.75B	128K	🤗 HuggingFace

Evaluation

For a comprehensive evaluation of the quality of our reasoning models, we implemented automatic benchmarks to assess their performance including math, code and science. The results indicate Ring-mini-2.0 achieves comparable performace with Ring-lite-2507 while activating only half parameters.

Quickstart

🤗 Hugging Face Transformers

Here is a code snippet to show you how to use the chat model with transformers:

from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "inclusionAI/Ring-mini-2.0"

model = AutoModelForCausalLM.from_pretrained(
    model_name,
    torch_dtype="auto",
    device_map="auto",
    trust_remote_code=True
)
tokenizer = AutoTokenizer.from_pretrained(model_name)

prompt = "Give me a short introduction to large language models."
messages = [
    {"role": "system", "content": "You are Ring, an assistant created by inclusionAI"},
    {"role": "user", "content": prompt}
]
text = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True,
    enable_thinking=True
)
model_inputs = tokenizer([text], return_tensors="pt", return_token_type_ids=False).to(model.device)

generated_ids = model.generate(
    **model_inputs,
    max_new_tokens=8192
)
generated_ids = [
    output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
]

response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]

Deployment

Please refer to GitHub

License

This code repository is licensed under the MIT License.

Citation

TODO