--- license: apache-2.0 base_model: google/gemma-2b model-index: - name: Octopus-V2-2B results: [] tags: - function calling - on-device language model - android inference: false language: - en --- # Octopus V2: On-device language model for super agent
- Nexa AI Product - ArXiv
## Introducing Octopus-V2-2B Octopus-V2-2B, an advanced open-source language model with 2 billion parameters, represents Nexa AI's research breakthrough in the application of large language models (LLMs) for function calling, specifically tailored for Android APIs. Unlike Retrieval-Augmented Generation (RAG) methods, which require detailed descriptions of potential function arguments—sometimes needing up to tens of thousands of input tokens—Octopus-V2-2B introduces a unique **functional token** strategy for both its training and inference stages. This approach not only allows it to achieve performance levels comparable to GPT-4 but also significantly enhances its inference speed beyond that of RAG-based methods, making it especially beneficial for edge computing devices. 📱 **On-device Applications**: Octopus-V2-2B is engineered to operate seamlessly on Android devices, extending its utility across a wide range of applications, from Android system management to the orchestration of multiple devices. Further demonstrations of its capabilities are available on the [Nexa AI Research Page](https://nexaai.github.io/octopus), showcasing its adaptability and potential for on-device integration. 🚀 **Inference Speed**: When benchmarked, Octopus-V2-2B demonstrates a remarkable inference speed, outperforming the combination of "Llama7B + RAG solution" by a factor of 36X on a single A100 GPU. Furthermore, compared to GPT-4-turbo (gpt-4-0125-preview), which relies on clusters A100/H100 GPUs, Octopus-V2-2B is 168% faster. This efficiency is attributed to our **functional token** design. 🐙 **Accuracy**: Octopus-V2-2B not only excels in speed but also in accuracy, surpassing the "Llama7B + RAG solution" in function call accuracy by 31%. It achieves a function call accuracy comparable to GPT-4 and RAG + GPT-3.5, with scores ranging between 98% and 100% across benchmark datasets. 💪 **Function Calling Capabilities**: Octopus-V2-2B is capable of generating individual, nested, and parallel function calls across a variety of complex scenarios. ## Example Use Cases You can run the model on a GPU using the following code. ```python from gemma.modeling_gemma import GemmaForCausalLM from transformers import AutoTokenizer import torch import time def inference(input_text): start_time = time.time() input_ids = tokenizer(input_text, return_tensors="pt").to(model.device) input_length = input_ids["input_ids"].shape[1] outputs = model.generate( input_ids=input_ids["input_ids"], max_length=1024, do_sample=False) generated_sequence = outputs[:, input_length:].tolist() res = tokenizer.decode(generated_sequence[0]) end_time = time.time() return {"output": res, "latency": end_time - start_time} model_id = "NexaAIDev/android_API_10k_data" tokenizer = AutoTokenizer.from_pretrained(model_id) model = GemmaForCausalLM.from_pretrained( model_id, torch_dtype=torch.bfloat16, device_map="auto" ) input_text = "Take a selfie for me with front camera" nexa_query = f"Below is the query from the users, please call the correct function and generate the parameters to call the function.\n\nQuery: {input_text} \n\nResponse:" start_time = time.time() print("nexa model result:\n", inference(nexa_query)) print("latency:", time.time() - start_time," s") ``` ## Evaluation ## Training Data We wrote 20 Android API descriptions to used to train the models, see [this file](android_functions.txt) for details. The Android API implementations for our demos, and our training data will be published later. Below is one Android API description example ``` def get_trending_news(category=None, region='US', language='en', max_results=5): """ Fetches trending news articles based on category, region, and language. Parameters: - category (str, optional): News category to filter by, by default use None for all categories. Optional to provide. - region (str, optional): ISO 3166-1 alpha-2 country code for region-specific news, by default, uses 'US'. Optional to provide. - language (str, optional): ISO 639-1 language code for article language, by default uses 'en'. Optional to provide. - max_results (int, optional): Maximum number of articles to return, by default, uses 5. Optional to provide. Returns: - list[str]: A list of strings, each representing an article. Each string contains the article's heading and URL. """ ``` ## License This model was trained on commercially viable data and is under the [Nexa AI community disclaimer](https://www.nexa4ai.com/disclaimer). ## References We thank the Google Gemma team for their amazing models! ``` @misc{gemma-2023-open-models, author = {{Gemma Team, Google DeepMind}}, title = {Gemma: Open Models Based on Gemini Research and Technology}, url = {https://goo.gle/GemmaReport}, year = {2023}, } ``` ## Citation ``` @misc{chen2024octopus, title={Octopus v2: On-device language model for super agent}, author={Wei Chen and Zhiyuan Li}, year={2024}, eprint={2404.01744}, archivePrefix={arXiv}, primaryClass={cs.CL} } ``` ## Contact Please [contact us](mailto:alexchen@nexa4ai.com) to reach out for any issues and comments!