ใICLR 2024๐ฅใ Extending Video-Language Pretraining to N-modality by Language-based Semantic Alignment
Video-LLaVA: Learning United Visual Representation by Alignment Before Projection
Bin Lin, Yang Ye, Bin Zhu, Jiaxi Cui, Munan Ning, Peng Jin, Li Yuan
MoE-LLaVA: Mixture of Experts for Large Vision-Language Models
Bin Lin, Zhenyu Tang, Yang Ye, Jiaxi Cui, Bin Zhu, Peng Jin, Junwu Zhang, Munan Ning, Li Yuan
Video-Bench: A Comprehensive Benchmark and Toolkit for Evaluating Video-based Large Language Models
Munan Ning, Bin Zhu, Yujia Xie, Bin Lin, Jiaxi Cui, Lu Yuan, Dongdong Chen, Li Yuan
LanguageBind is a language-centric multimodal pretraining approach, taking the language as the bind across different modalities because the language modality is well-explored and contains rich semantics.
We propose VIDAL-10M, 10 Million data with Video, Infrared, Depth, Audio and their corresponding Language, which greatly expands the data beyond visual modalities.
We make multi-view enhancements to language. We produce multi-view description that combines meta-data, spatial, and temporal to greatly enhance the semantic information of the language. In addition we further enhance the language with ChatGPT to create a good semantic space for each modality aligned language.
python gradio_app.py
LanguageBind achieves state-of-the-art (SOTA) performance on four datasets, * donates the results of full tuning.
Video-Language, Infrared-Language, Depth-Language, and Audio-Language zero-shot classification, * donates the results of full tuning.
We report text-to-audio results for retrieval, * donates the results of full tuning.
git clone https://github.com/PKU-YuanGroup/LanguageBind
cd LanguageBind
pip install torch==1.13.1+cu116 torchvision==0.14.1+cu116 torchaudio==0.13.1 --extra-index-url https://download.pytorch.org/whl/cu116
pip install -r requirements.txt
The names in the table represent different encoder models. For example, LanguageBind/LanguageBind_Video_FT
represents the fully fine-tuned version, while LanguageBind/LanguageBind_Video
represents the LoRA-tuned version.
You can freely replace them in the recommended API usage. We recommend using the fully fine-tuned version, as it offers stronger performance.
Modality | LoRA tuning | Fine-tuning |
---|---|---|
Video | LanguageBind_Video | LanguageBind_Video_FT |
Audio | LanguageBind_Audio | LanguageBind_Audio_FT |
Depth | LanguageBind_Depth | - |
Thermal | LanguageBind_Thermal | - |
Version | Tuning | Model size | Num_frames | HF Link | MSR-VTT | DiDeMo | ActivityNet | MSVD |
---|---|---|---|---|---|---|---|---|
LanguageBind_Video | LoRA | Large | 8 | Link | 42.6 | 37.8 | 35.1 | 52.2 |
LanguageBind_Video_FT | Full-tuning | Large | 8 | Link | 42.7 | 38.1 | 36.9 | 53.5 |
LanguageBind_Video_V1.5_FT | Full-tuning | Large | 8 | Link | 42.8 | 39.7 | 38.4 | 54.1 |
LanguageBind_Video_V1.5_FT | Full-tuning | Large | 12 | Coming soon | ||||
LanguageBind_Video_Huge_V1.5_FT | Full-tuning | Huge | 8 | Link | 44.8 | 39.9 | 41.0 | 53.7 |
LanguageBind_Video_Huge_V1.5_FT | Full-tuning | Huge | 12 | Coming soon |
We open source all modalities preprocessing code. If you want to load the model (e.g. LanguageBind/LanguageBind_Thermal
) from the model hub on Huggingface or on local, you can use the following code snippets!
We have provided some sample datasets in assets to quickly see how languagebind works.
import torch
from languagebind import LanguageBind, to_device, transform_dict, LanguageBindImageTokenizer
if __name__ == '__main__':
device = 'cuda:0'
device = torch.device(device)
clip_type = {
'video': 'LanguageBind_Video_FT', # also LanguageBind_Video
'audio': 'LanguageBind_Audio_FT', # also LanguageBind_Audio
'thermal': 'LanguageBind_Thermal',
'image': 'LanguageBind_Image',
'depth': 'LanguageBind_Depth',
}
model = LanguageBind(clip_type=clip_type, cache_dir='./cache_dir')
model = model.to(device)
model.eval()
pretrained_ckpt = f'lb203/LanguageBind_Image'
tokenizer = LanguageBindImageTokenizer.from_pretrained(pretrained_ckpt, cache_dir='./cache_dir/tokenizer_cache_dir')
modality_transform = {c: transform_dict[c](model.modality_config[c]) for c in clip_type.keys()}
image = ['assets/image/0.jpg', 'assets/image/1.jpg']
audio = ['assets/audio/0.wav', 'assets/audio/1.wav']
video = ['assets/video/0.mp4', 'assets/video/1.mp4']
depth = ['assets/depth/0.png', 'assets/depth/1.png']
thermal = ['assets/thermal/0.jpg', 'assets/thermal/1.jpg']
language = ["Training a parakeet to climb up a ladder.", 'A lion climbing a tree to catch a monkey.']
inputs = {
'image': to_device(modality_transform['image'](image), device),
'video': to_device(modality_transform['video'](video), device),
'audio': to_device(modality_transform['audio'](audio), device),
'depth': to_device(modality_transform['depth'](depth), device),
'thermal': to_device(modality_transform['thermal'](thermal), device),
}
inputs['language'] = to_device(tokenizer(language, max_length=77, padding='max_length',
truncation=True, return_tensors='pt'), device)
with torch.no_grad():
embeddings = model(inputs)
print("Video x Text: \n",
torch.softmax(embeddings['video'] @ embeddings['language'].T, dim=-1).detach().cpu().numpy())
print("Image x Text: \n",
torch.softmax(embeddings['image'] @ embeddings['language'].T, dim=-1).detach().cpu().numpy())
print("Depth x Text: \n",
torch.softmax(embeddings['depth'] @ embeddings['language'].T, dim=-1).detach().cpu().numpy())
print("Audio x Text: \n",
torch.softmax(embeddings['audio'] @ embeddings['language'].T, dim=-1).detach().cpu().numpy())
print("Thermal x Text: \n",
torch.softmax(embeddings['thermal'] @ embeddings['language'].T, dim=-1).detach().cpu().numpy())
Then returns the following result.
Video x Text:
[[9.9989331e-01 1.0667283e-04]
[1.3255903e-03 9.9867439e-01]]
Image x Text:
[[9.9990666e-01 9.3292067e-05]
[4.6132666e-08 1.0000000e+00]]
Depth x Text:
[[0.9954276 0.00457235]
[0.12042473 0.8795753 ]]
Audio x Text:
[[0.97634876 0.02365119]
[0.02917843 0.97082156]]
Thermal x Text:
[[0.9482511 0.0517489 ]
[0.48746133 0.5125386 ]]
Since languagebind binds each modality together, we also found the emergency zero-shot. It's very simple to use.
print("Video x Audio: \n", torch.softmax(embeddings['video'] @ embeddings['audio'].T, dim=-1).detach().cpu().numpy())
print("Image x Depth: \n", torch.softmax(embeddings['image'] @ embeddings['depth'].T, dim=-1).detach().cpu().numpy())
print("Image x Thermal: \n", torch.softmax(embeddings['image'] @ embeddings['thermal'].T, dim=-1).detach().cpu().numpy())
Then, you will get:
Video x Audio:
[[1.0000000e+00 0.0000000e+00]
[3.1150486e-32 1.0000000e+00]]
Image x Depth:
[[1. 0.]
[0. 1.]]
Image x Thermal:
[[1. 0.]
[0. 1.]]
Additionally, LanguageBind can be disassembled into different branches to handle different tasks. Note that we do not train Image, which just initialize from OpenCLIP.
import torch
from languagebind import LanguageBindThermal, LanguageBindThermalTokenizer, LanguageBindThermalProcessor
pretrained_ckpt = 'LanguageBind/LanguageBind_Thermal'
model = LanguageBindThermal.from_pretrained(pretrained_ckpt, cache_dir='./cache_dir')
tokenizer = LanguageBindThermalTokenizer.from_pretrained(pretrained_ckpt, cache_dir='./cache_dir')
thermal_process = LanguageBindThermalProcessor(model.config, tokenizer)
model.eval()
data = thermal_process([r"your/thermal.jpg"], ['your text'], return_tensors='pt')
with torch.no_grad():
out = model(**data)
print(out.text_embeds @ out.image_embeds.T)
import torch
from languagebind import LanguageBindDepth, LanguageBindDepthTokenizer, LanguageBindDepthProcessor
pretrained_ckpt = 'LanguageBind/LanguageBind_Depth'
model = LanguageBindDepth.from_pretrained(pretrained_ckpt, cache_dir='./cache_dir')
tokenizer = LanguageBindDepthTokenizer.from_pretrained(pretrained_ckpt, cache_dir='./cache_dir')
depth_process = LanguageBindDepthProcessor(model.config, tokenizer)
model.eval()
data = depth_process([r"your/depth.png"], ['your text.'], return_tensors='pt')
with torch.no_grad():
out = model(**data)
print(out.text_embeds @ out.image_embeds.T)
import torch
from languagebind import LanguageBindVideo, LanguageBindVideoTokenizer, LanguageBindVideoProcessor
pretrained_ckpt = 'LanguageBind/LanguageBind_Video_FT' # also 'LanguageBind/LanguageBind_Video'
model = LanguageBindVideo.from_pretrained(pretrained_ckpt, cache_dir='./cache_dir')
tokenizer = LanguageBindVideoTokenizer.from_pretrained(pretrained_ckpt, cache_dir='./cache_dir')
video_process = LanguageBindVideoProcessor(model.config, tokenizer)
model.eval()
data = video_process(["your/video.mp4"], ['your text.'], return_tensors='pt')
with torch.no_grad():
out = model(**data)
print(out.text_embeds @ out.image_embeds.T)
import torch
from languagebind import LanguageBindAudio, LanguageBindAudioTokenizer, LanguageBindAudioProcessor
pretrained_ckpt = 'LanguageBind/LanguageBind_Audio_FT' # also 'LanguageBind/LanguageBind_Audio'
model = LanguageBindAudio.from_pretrained(pretrained_ckpt, cache_dir='./cache_dir')
tokenizer = LanguageBindAudioTokenizer.from_pretrained(pretrained_ckpt, cache_dir='./cache_dir')
audio_process = LanguageBindAudioProcessor(model.config, tokenizer)
model.eval()
data = audio_process([r"your/audio.wav"], ['your audio.'], return_tensors='pt')
with torch.no_grad():
out = model(**data)
print(out.text_embeds @ out.image_embeds.T)
Note that our image encoder is the same as OpenCLIP. Not as fine-tuned as other modalities.
import torch
from languagebind import LanguageBindImage, LanguageBindImageTokenizer, LanguageBindImageProcessor
pretrained_ckpt = 'LanguageBind/LanguageBind_Image'
model = LanguageBindImage.from_pretrained(pretrained_ckpt, cache_dir='./cache_dir')
tokenizer = LanguageBindImageTokenizer.from_pretrained(pretrained_ckpt, cache_dir='./cache_dir')
image_process = LanguageBindImageProcessor(model.config, tokenizer)
model.eval()
data = image_process([r"your/image.jpg"], ['your text.'], return_tensors='pt')
with torch.no_grad():
out = model(**data)
print(out.text_embeds @ out.image_embeds.T)
The datasets is in DATASETS.md.
The training & validating instruction is in TRAIN_AND_VALIDATE.md.
If you find our paper and code useful in your research, please consider giving a star :star: and citation :pencil:.
@misc{zhu2023languagebind,
title={LanguageBind: Extending Video-Language Pretraining to N-modality by Language-based Semantic Alignment},
author={Bin Zhu and Bin Lin and Munan Ning and Yang Yan and Jiaxi Cui and Wang HongFa and Yatian Pang and Wenhao Jiang and Junwu Zhang and Zongwei Li and Cai Wan Zhang and Zhifeng Li and Wei Liu and Li Yuan},
year={2023},
eprint={2310.01852},
archivePrefix={arXiv},
primaryClass={cs.CV}
}