openbmb / MiniCPM-o-2_6

A GPT-4o Level MLLM for Vision, Speech and Multimodal Live Streaming on Your Phone

GitHub | Online Demo

MiniCPM-o 2.6

MiniCPM-o 2.6 is the latest and most capable model in the MiniCPM-o series. The model is built in an end-to-end fashion based on SigLip-400M, Whisper-medium-300M, ChatTTS-200M, and Qwen2.5-7B with a total of 8B parameters. It exhibits a significant performance improvement over MiniCPM-V 2.6, and introduces new features for realtime speech conversation and multimodal live streaming. Notable features of MiniCPM-o 2.6 include:

• 🔥 Leading Visual Capability. MiniCPM-o 2.6 achieves an average score of 70.2 on OpenCompass, a comprehensive evaluation over 8 popular benchmarks. With only 8B parameters, it surpasses widely used proprietary models like GPT-4o-202405, Gemini 1.5 Pro, and Claude 3.5 Sonnet for single image understanding. It also outperforms GPT-4V and Claude 3.5 Sonnet in mutli-image and video understanding, and shows promising in-context learning capability.

• 🎙 State-of-the-art Speech Capability. MiniCPM-o 2.6 supports bilingual realtime speech conversation with configurable voices in English and Chinese. It outperforms GPT-4o-realtime on audio understanding tasks such as ASR and STT translation, and shows state-of-the-art performance on speech conversation in both semantic and acoustic evaluations in the open-source community . It also allows for fun features such as emotion/speed/style control, end-to-end voice cloning, role play, etc.

• 🎬 Strong Multimodal Live Streaming Capability. As a new feature, MiniCPM-o 2.6 can accept continous video and audio streams independent of user queries, and support realtime speech interaction . It outperforms GPT-4o-realtime and Claude 3.5 Sonnet and shows state-of-art performance in open-source community on StreamingBench , a comprehensive benchmark for real-time video understanding, omni-source (video & audio) understanding , and multimodal contextual understanding.

• 💪 Strong OCR Capability and Others. Advancing popular visual capabilites from MiniCPM-V series, MiniCPM-o 2.6 can process images with any aspect ratio and up to 1.8 million pixels (e.g., 1344x1344). It achieves state-of-the-art performance on OCRBench for models under 25B, surpassing proprietary models such as GPT-4o-202405 . Based on the the latest RLAIF-V and VisCPM techniques, it features trustworthy behaviors , outperforming GPT-4o and Claude 3.5 Sonnet on MMHal-Bench, and supports multilingual capabilities on more than 30 languages.

• 🚀 Superior Efficiency. In addition to its friendly size, MiniCPM-o 2.6 also shows state-of-the-art token density (i.e., number of pixels encoded into each visual token). It produces only 640 tokens when processing a 1.8M pixel image, which is 75% fewer than most models . This directly improves the inference speed, first-token latency, memory usage, and power consumption. As a result, MiniCPM-o 2.6 can efficiently support multimodal live streaming on end-side devices such as iPad.

• 💫 Easy Usage. MiniCPM-o 2.6 can be easily used in various ways: (1) llama.cpp support for efficient CPU inference on local devices, (2) int4 and GGUF format quantized models in 16 sizes, (3) vLLM support for high-throughput and memory-efficient inference, (4) fine-tuning on new domains and tasks with LLaMA-Factory , (5) quick local WebUI demo setup with Gradio , and (6) online web demo on CN server and US server.

Model Architecture.

• End-to-end Omni-modal Architecture. Different modality encoder/decoders are connected and trained in an end-to-end fashion to fully exploit rich multimodal knowledge.
• Omni-modal Live Streaming Mechanism. (1) We change the offline modality encoder/decoders into online ones for streaminig inputs/outputs. (2) We devise a time-division multiplexing (TDM) mechanism for omni-modality streaminig processing in the LLM backbone. It divides parallel omni-modality streams into sequential info within small periodic time slices.
• Configurable Speech Modeling Design. We devise a multimodal system prompt, including traditional text system prompt, and a new audio system prompt to determine the assistant voice . This enables flexible voice configurations in inference time, and also facilitates end-to-end voice cloning and description-based voice creation.

Evaluation

Examples

We deploy MiniCPM-o 2.6 on end devices. The demo video is the raw screen recording on a iPad Pro without edition.

Online Demo

Click here to try the online demo of MiniCPM-o 2.6 .

Usage

Inference using Huggingface transformers on NVIDIA GPUs. Requirements tested on python 3.10：

Pillow==10.1.0
torch==2.2.0
torchaudio==2.2.0
torchvision==0.17.0
transformers==4.44.2
librosa==0.9.0
soundfile==0.12.1
vector-quantize-pytorch==1.18.5
vocos==0.1.0
decord
moviepy

Model initialization

import torch
from PIL import Image
from transformers import AutoModel, AutoTokenizer

# load omni model default, the default init_vision/init_audio/init_tts is True
# if load vision-only model, please set init_audio=False and init_tts=False
# if load audio-only model, please set init_vision=False
model = AutoModel.from_pretrained(
    'openbmb/MiniCPM-o-2_6',
    trust_remote_code=True,
    attn_implementation='sdpa', # sdpa or flash_attention_2
    torch_dtype=torch.bfloat16,
    init_vision=True,
    init_audio=True,
    init_tts=True
)


model = model.eval().cuda()
tokenizer = AutoTokenizer.from_pretrained('openbmb/MiniCPM-o-2_6', trust_remote_code=True)

# In addition to vision-only mode, tts processor and vocos also needs to be initialized
model.init_tts()
model.tts.float()

Omni mode

we provide two inference modes: chat and streaming

Chat inference

import math
import numpy as np
from PIL import Image
from moviepy.editor import VideoFileClip
import tempfile
import librosa
import soundfile as sf

def get_video_chunk_content(video_path, flatten=True):
    video = VideoFileClip(video_path)
    print('video_duration:', video.duration)
    
    with tempfile.NamedTemporaryFile(suffix=".wav", delete=True) as temp_audio_file:
        temp_audio_file_path = temp_audio_file.name
        video.audio.write_audiofile(temp_audio_file_path, codec="pcm_s16le", fps=16000)
        audio_np, sr = librosa.load(temp_audio_file_path, sr=16000, mono=True)
    num_units = math.ceil(video.duration)
    
    # 1 frame + 1s audio chunk
    contents= []
    for i in range(num_units):
        frame = video.get_frame(i+1)
        image = Image.fromarray((frame).astype(np.uint8))
        audio = audio_np[sr*i:sr*(i+1)]
        if flatten:
            contents.extend(["<unit>", image, audio])
        else:
            contents.append(["<unit>", image, audio])
    
    return contents

video_path="/path/to/video"
# if use voice clone prompt, please set ref_audio
ref_audio_path = 'assets/demo.wav'
ref_audio, _ = librosa.load(ref_audio_path, sr=16000, mono=True)
sys_msg = model.get_sys_prompt(ref_audio=ref_audio, mode='omni', language='en')
# or use default prompt
# sys_msg = model.get_sys_prompt(mode='omni', language='en')

contents = get_video_chunk_content(video_path)
msg = {"role":"user", "content": contents}
msgs = [sys_msg, msg]

# please set generate_audio=True and output_audio_path to save the tts result
generate_audio = True
output_audio_path = 'output.wav'

res = model.chat(
    msgs=msgs,
    tokenizer=tokenizer,
    sampling=True,
    temperature=0.5,
    max_new_tokens=4096,
    omni_input=True, # please set omni_input=True when omni inference
    use_tts_template=True,
    generate_audio=generate_audio,
    output_audio_path=output_audio_path,
    max_slice_nums=1,
    use_image_id=False,
    return_dict=True
)
print(res)

Streaming inference

# a new conversation need reset session first, it will reset the kv-cache
model.reset_session()

contents = get_video_chunk_content(video_path, flatten=False)
session_id = '123'
generate_audio = True

# 1. prefill system prompt
res = model.streaming_prefill(
    session_id=session_id,
    msgs=[sys_msg], 
    tokenizer=tokenizer
)

# 2. prefill video/audio chunks
for content in contents:
    msgs = [{"role":"user", "content": content}]
    res = model.streaming_prefill(
        session_id=session_id,
        msgs=msgs, 
        tokenizer=tokenizer
    )

# 3. generate
res = model.streaming_generate(
    session_id=session_id,
    tokenizer=tokenizer,
    temperature=0.5,
    generate_audio=generate_audio
)

audios = []
text = ""

if generate_audio:
    for r in res:
        audio_wav = r.audio_wav
        sampling_rate = r.sampling_rate
        txt = r.text

        audios.append(audio_wav)
        text += txt
        
    res = np.concatenate(audios)
    sf.write("output.wav", res, samplerate=sampling_rate)
    print("text:", text)
    print("audio saved to output.wav")
else:
    for r in res:
        text += r['text']
    print("text:", text)

Audio-Only mode

Mimick

Mimick task reflects a model's end-to-end speech modeling capability. The model takes audio input, and outputs an ASR transcription and subsequently reconstructs the original audio with high similarity. The higher the similarity between the reconstructed audio and the original audio, the stronger the model's foundational capability in end-to-end speech modeling.

mimick_prompt = "Please repeat each user's speech, including voice style and speech content."
audio_input, _ = librosa.load('xxx.wav', sr=16000, mono=True)
msgs = [{'role': 'user', 'content': [mimick_prompt,audio_input]}]

res = model.chat(
    msgs=msgs,
    tokenizer=tokenizer,
    sampling=True,
    max_new_tokens=128,
    use_tts_template=True,
    temperature=0.3,
    generate_audio=True,
    output_audio_path='output.wav', # save the tts result to output_audio_path
)

General Speech Conversation with Configurable Voices

ref_audio, _ = librosa.load('assets/demo.wav', sr=16000, mono=True) # load the reference audio

# Choose the mode you want to use
# Audio RolePlay:  # With this mode, model will role-play the character based on the audio prompt. (More human-like conversation but unstable)
# sys_prompt = model.get_sys_prompt(ref_audio=ref_audio, mode='audio_roleplay', language='en')
# user_question = {'role': 'user', 'content': [librosa.load('xxx.wav', sr=16000, mono=True)[0]]}

Audio Assistant: # With this mode, model will speak with the voice in ref_audio as a AI assistant. (Stable and more suitable for general conversation)
sys_prompt = model.get_sys_prompt(ref_audio=ref_audio, mode='audio_assistant', language='en') 
user_question = {'role': 'user', 'content': [librosa.load('xxx.wav', sr=16000, mono=True)[0]]} # Try to ask something by recording it in 'xxx.wav'!!!

msgs = [sys_prompt, user_question]
# round one
res = model.chat(
    msgs=msgs,
    tokenizer=tokenizer,
    sampling=True,
    max_new_tokens=128,
    use_tts_template=True,
    generate_audio=True,
    temperature=0.3,
    output_audio_path='result.wav',
)

# round two
history = msgs.append({'role': 'assistant', 'content': res})
user_question = {'role': 'user', 'content': [librosa.load('xxx.wav', sr=16000, mono=True)[0]]}
msgs = history.append(user_question)
res = model.chat(
    msgs=msgs,
    tokenizer=tokenizer,
    sampling=True,
    max_new_tokens=128,
    use_tts_template=True,
    generate_audio=True,
    temperature=0.3,
    output_audio_path='result_round_2.wav',
)
print(res)

Addressing various audio tasks

'''
Audio Understanding Task Prompt:
Speech:
    ASR with ZH(same as AST en2zh): 请仔细听这段音频片段，并将其内容逐字记录。
    ASR with EN(same as AST zh2en): Please listen to the audio snippet carefully and transcribe the content.
    Speaker Analysis: Based on the speaker's content, speculate on their gender, condition, age range, and health status.
General Audio:
    Audio Caption: Summarize the main content of the audio.
    Sound Scene Tagging: Utilize one keyword to convey the audio's content or the associated scene.
'''
task_prompt = "" # Choose the task prompt above
audio_input, _ = librosa.load('xxx.wav', sr=16000, mono=True)

msgs = [{'role': 'user', 'content': [task_prompt,audio_input]}]

res = model.chat(
    msgs=msgs,
    tokenizer=tokenizer,
    sampling=True,
    max_new_tokens=128,
    use_tts_template=True,
    generate_audio=True,
    temperature=0.3,
    output_audio_path='result.wav',
)
print(res)

'''
Speech Generation Task Prompt:
    Human Instruction-to-Speech: see https://voxinstruct.github.io/VoxInstruct/
    Example:
        # 在新闻中，一个年轻男性兴致勃勃地说：“祝福亲爱的祖国母亲美丽富强！”他用低音调和低音量，慢慢地说出了这句话。
        # Delighting in a surprised tone, an adult male with low pitch and low volume comments:"One even gave my little dog a biscuit" This dialogue takes place at a leisurely pace, delivering a sense of excitement and surprise in the context. 

    Voice Cloning or Voice Conversion: With this mode, model will act like a TTS model. 
'''
# Human Instruction-to-Speech:
task_prompt = '' #Try to make some Human Instruction-to-Speech prompt (Voice Creation)
msgs = [{'role': 'user', 'content': [task_prompt]}] # you can also try to ask the same audio question

# Voice Cloning mode: 
# sys_prompt = model.get_sys_prompt(ref_audio=ref_audio, mode='voice_cloning', language='en')
# text_prompt = f"Please read the text below."
# user_question = {'role': 'user', 'content': [text_prompt, "content that you want to read"]} # using same voice in sys_prompt to read the text. (Voice Cloning)
# user_question = {'role': 'user', 'content': [text_prompt, librosa.load('xxx.wav', sr=16000, mono=True)[0]]} # using same voice in sys_prompt to read 'xxx.wav'. (Voice Conversion)
# msgs = [sys_prompt, user_question]

res = model.chat(
    msgs=msgs,
    tokenizer=tokenizer,
    sampling=True,
    max_new_tokens=128,
    use_tts_template=True,
    generate_audio=True,
    temperature=0.3,
    output_audio_path='result.wav',
)

Vision-Only mode

MiniCPM-o-2_6 has the same inference methods as MiniCPM-V-2_6

Chat with single image

# test.py
image = Image.open('xx.jpg').convert('RGB')
question = 'What is in the image?'
msgs = [{'role': 'user', 'content': [image, question]}]
res = model.chat(
    image=None,
    msgs=msgs,
    tokenizer=tokenizer
)
print(res)

## if you want to use streaming, please make sure sampling=True and stream=True
## the model.chat will return a generator
res = model.chat(
    msgs=msgs,
    tokenizer=tokenizer,
    sampling=True,
    stream=True
)
generated_text = ""
for new_text in res:
    generated_text += new_text
    print(new_text, flush=True, end='')

Chat with multiple images

image1 = Image.open('image1.jpg').convert('RGB')
image2 = Image.open('image2.jpg').convert('RGB')
question = 'Compare image 1 and image 2, tell me about the differences between image 1 and image 2.'
msgs = [{'role': 'user', 'content': [image1, image2, question]}]
answer = model.chat(
    msgs=msgs,
    tokenizer=tokenizer
)
print(answer)

In-context few-shot learning

question = "production date" 
image1 = Image.open('example1.jpg').convert('RGB')
answer1 = "2023.08.04"
image2 = Image.open('example2.jpg').convert('RGB')
answer2 = "2007.04.24"
image_test = Image.open('test.jpg').convert('RGB')
msgs = [
    {'role': 'user', 'content': [image1, question]}, {'role': 'assistant', 'content': [answer1]},
    {'role': 'user', 'content': [image2, question]}, {'role': 'assistant', 'content': [answer2]},
    {'role': 'user', 'content': [image_test, question]}
]
answer = model.chat(
    msgs=msgs,
    tokenizer=tokenizer
)
print(answer)

Chat with video

MAX_NUM_FRAMES=64 # if cuda OOM set a smaller number
def encode_video(video_path):
    def uniform_sample(l, n):
        gap = len(l) / n
        idxs = [int(i * gap + gap / 2) for i in range(n)]
        return [l[i] for i in idxs]
    vr = VideoReader(video_path, ctx=cpu(0))
    sample_fps = round(vr.get_avg_fps() / 1)  # FPS
    frame_idx = [i for i in range(0, len(vr), sample_fps)]
    if len(frame_idx) > MAX_NUM_FRAMES:
        frame_idx = uniform_sample(frame_idx, MAX_NUM_FRAMES)
    frames = vr.get_batch(frame_idx).asnumpy()
    frames = [Image.fromarray(v.astype('uint8')) for v in frames]
    print('num frames:', len(frames))
    return frames
video_path ="video_test.mp4"
frames = encode_video(video_path)
question = "Describe the video"
msgs = [
    {'role': 'user', 'content': frames + [question]}, 
]
# Set decode params for video
params={}
params["use_image_id"] = False
params["max_slice_nums"] = 2 # use 1 if cuda OOM and video resolution >  448*448
answer = model.chat(
    msgs=msgs,
    tokenizer=tokenizer,
    **params
)
print(answer)

Please look at GitHub for more detail about usage.

Inference with llama.cpp

MiniCPM-o 2.6 can run with llama.cpp. See our fork of llama.cpp for more detail.

Int4 quantized version

Download the int4 quantized version for lower GPU memory (7GB) usage: MiniCPM-o-2_6-int4 .

License

Model License

• The code in this repo is released under the Apache-2.0 License.
• The usage of MiniCPM-o and MiniCPM-V series model weights must strictly follow MiniCPM Model License.md .
• The models and weights of MiniCPM are completely free for academic research. After filling out a "questionnaire" for registration, MiniCPM-o 2.6 weights are also available for free commercial use.

Statement

• As an LMM, MiniCPM-o 2.6 generates contents by learning a large mount of multimodal corpora, but it cannot comprehend, express personal opinions or make value judgement. Anything generated by MiniCPM-o 2.6 does not represent the views and positions of the model developers
• We will not be liable for any problems arising from the use of the MinCPM-V models, including but not limited to data security issues, risk of public opinion, or any risks and problems arising from the misdirection, misuse, dissemination or misuse of the model.

Key Techniques and Other Multimodal Projects

👏 Welcome to explore key techniques of MiniCPM-o 2.6 and other multimodal projects of our team:

VisCPM | RLHF-V | LLaVA-UHD | RLAIF-V

Citation

If you find our work helpful, please consider citing our papers 📝 and liking this project ❤️！

@article{yao2024minicpm,
  title={MiniCPM-V: A GPT-4V Level MLLM on Your Phone},
  author={Yao, Yuan and Yu, Tianyu and Zhang, Ao and Wang, Chongyi and Cui, Junbo and Zhu, Hongji and Cai, Tianchi and Li, Haoyu and Zhao, Weilin and He, Zhihui and others},
  journal={arXiv preprint arXiv:2408.01800},
  year={2024}
}