Rwkv7 1.5B World

🚀 rwkv7-1.5B-world

This is an RWKV-7 model in flash-linear attention format, capable of handling multiple languages such as English, Chinese, Japanese, etc.

🚀 Quick Start

Before using this model, you need to install flash-linear-attention and the latest version of transformers:

pip install git+https://github.com/fla-org/flash-linear-attention
pip install 'transformers>=4.48.0'

✨ Features

• Multilingual Support: Supports multiple languages including English, Chinese, Japanese, Korean, French, Arabic, Spanish, and Portuguese.
• Flash-Linear Attention Format: Utilizes the flash-linear attention format for better performance.

📦 Installation

pip install git+https://github.com/fla-org/flash-linear-attention
pip install 'transformers>=4.48.0'

💻 Usage Examples

Basic Usage

from transformers import AutoModelForCausalLM, AutoTokenizer
model = AutoModelForCausalLM.from_pretrained('fla-hub/rwkv7-1.5B-world', trust_remote_code=True)
tokenizer = AutoTokenizer.from_pretrained('fla-hub/rwkv7-1.5B-world', trust_remote_code=True)

model = model.cuda() # Supported on Nvidia/AMD/Intel eg. model.xpu()
prompt = "What is a large language model?"
messages = [
    {"role": "user", "content": prompt}
]
text = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True
)

model_inputs = tokenizer([text], return_tensors="pt").to(model.device)

generated_ids = model.generate(
    **model_inputs,
    max_new_tokens=4096,
    do_sample=True,
    temperature=1.0,
    top_p=0.3,
    repetition_penalty=1.2
)
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=False)[0]
print(response)

📚 Documentation

Model Details

Model Description

• Developed by: Bo Peng, Yu Zhang, Songlin Yang, Ruichong Zhang
• Funded by: RWKV Project (Under LF AI & Data Foundation)
• Model type: RWKV7
• Language(s) (NLP): English, Chinese, Japanese, Korean, French, Arabic, Spanish, Portuguese
• License: Apache-2.0
• Parameter count: 1.52B
• Tokenizer: RWKV World tokenizer
• Vocabulary size: 65,536

Model Sources

• Repository: https://github.com/fla-org/flash-linear-attention ; https://github.com/BlinkDL/RWKV-LM
• Paper: https://huggingface.co/papers/2503.14456

Training Details

Training Data

This model is trained on the World v3 with a total of 3.119 trillion tokens.

Training Hyperparameters

• Training regime: bfloat16, lr 4e-4 to 1e-5 "delayed" cosine decay, wd 0.1 (with increasing batch sizes during the middle)
• Final Loss: 1.9965
• Token Count: 3.119 trillion

Evaluation

Metrics

lambada_openai:

• Before conversion: ppl 4.13 acc 69.4%
• After conversion: ppl 4.26 acc 68.8% (without apply temple)

FAQ

⚠️ Important Note

If the safetensors metadata is none, you need to upgrade transformers to >=4.48.0: pip install 'transformers>=4.48.0'

📄 License

This model is licensed under the Apache-2.0 license.