# image InternVL Family: Closing the Gap to Commercial Multimodal Models with Open-Source Suites โ€”โ€” A Pioneering Open-Source Alternative to GPT-4o [\[๐Ÿ†• Blog\]](https://internvl.github.io/blog/) [\[๐Ÿš€ InternVL2 Blog\]](https://internvl.github.io/blog/2024-07-02-InternVL-2.0/) [\[๐Ÿค— HF Chat Demo\]](https://huggingface.co/spaces/OpenGVLab/InternVL) [\[๐Ÿ—จ๏ธ Chat Demo\]](https://internvl.opengvlab.com/) [\[๐ŸŒ API\]](./document/How_to_use_InternVL_API.md) [\[๐Ÿš€ Quick Start\]](#quick-start-with-huggingface) [\[๐Ÿ“œ InternVL 1.0 Paper\]](https://arxiv.org/abs/2312.14238) [\[๐Ÿ“œ InternVL 1.5 Report\]](https://arxiv.org/abs/2404.16821) [\[๐Ÿ“– 1.0 ไธญๆ–‡่งฃ่ฏป\]](https://zhuanlan.zhihu.com/p/702946079) [\[๐Ÿ“– 1.5 ไธญๆ–‡่งฃ่ฏป\]](https://zhuanlan.zhihu.com/p/699439759) [\[๐Ÿ“– 2.0 ไธญๆ–‡่งฃ่ฏป\]](https://zhuanlan.zhihu.com/p/706547971) [Switch to the Chinese version (ๅˆ‡ๆข่‡ณไธญๆ–‡็‰ˆ)](/README_zh.md) OpenGVLab%2FInternVL | Trendshift image ![opencompass](https://github.com/user-attachments/assets/7ce93c05-84ae-4997-a480-53897d1d3a1c)
## News ๐Ÿš€๐Ÿš€๐Ÿš€ - `2024/07/18`: ๐Ÿ”ฅ๐Ÿ”ฅ InternVL2-40B achieved SOTA performance among open-source models on the [Video-MME](https://github.com/BradyFU/Video-MME) dataset, scoring 61.2 when inputting 16 frames and 64.4 when inputting 32 frames. It significantly outperforms other open-source models and is the closest open-source model to GPT-4o mini. - `2024/07/18`: ๐Ÿ”ฅ InternVL2-Pro achieved the SOTA performance on the [DocVQA](https://rrc.cvc.uab.es/?ch=17&com=evaluation&task=1) and [InfoVQA](https://rrc.cvc.uab.es/?ch=17&com=evaluation&task=3) benchmarks. - `2024/07/04`: ๐Ÿš€ We release the [InternVL2 series](https://huggingface.co/collections/OpenGVLab/internvl-20-667d3961ab5eb12c7ed1463e). InternVL2-Pro achieved a 62.0% accuracy on the MMMU benchmark, matching the performance of leading closed-source commercial models like GPT-4o. The free API of this model can be applied by filling ([application form](https://docs.google.com/forms/d/e/1FAIpQLSfMCzhPr1OOEKau_6jwTU0EiZMSFckDo-HMlc_hUudhF_97rw/viewform?usp=sf_link)) / ([็”ณ่ฏท่กจ](https://wj.qq.com/s2/14910502/25a4/)). Other models are available at [HF link](https://huggingface.co/collections/OpenGVLab/internvl-20-667d3961ab5eb12c7ed1463e). - `2024/06/19`: We propose Needle In A Multimodal Haystack ([MM-NIAH](https://github.com/OpenGVLab/MM-NIAH)), the first benchmark designed to systematically evaluate the capability of existing MLLMs to comprehend long multimodal documents. - `2024/05/30`: We release [ShareGPT-4o](https://sharegpt4o.github.io/), a large-scale dataset that we plan to open-source with 200K images, 10K videos, and 10K audios with detailed descriptions. - `2024/05/29`: We release the Mini-InternVL series, which includes two chat models: [Mini-InternVL-Chat-2B-V1-5](https://huggingface.co/OpenGVLab/Mini-InternVL-Chat-2B-V1-5) and [Mini-InternVL-Chat-4B-V1-5](https://huggingface.co/OpenGVLab/Mini-InternVL-Chat-4B-V1-5). These models achieve impressive performance with minimal size: the 2B model delivers 80% of the performance with only 8% of the model size, and the 4B model achieves 90% of the performance with just 16% of the model size. For more details, please check our [blog](https://internvl.github.io/blog/2024-05-25-Mini-InternVL-1.5/). - `2024/05/28`: Thanks to the [lmdeploy](https://github.com/InternLM/lmdeploy) team for providing AWQ quantization support. The 4-bit model is available at [OpenGVLab/InternVL-Chat-V1-5-AWQ](https://huggingface.co/OpenGVLab/InternVL-Chat-V1-5-AWQ). - `2024/05/13`: InternVL 1.0 can now be used as the [text encoder](https://huggingface.co/OpenGVLab/InternVL-14B-224px) for diffusion models to support multilingual generation natively in over 110 languages worldwide. See [MuLan](https://github.com/mulanai/MuLan) for more details. - `2024/04/18`: InternVL-Chat-V1-5 has been released at [HF link](https://huggingface.co/OpenGVLab/InternVL-Chat-V1-5), approaching the performance of GPT-4V and Gemini Pro on various benchmarks like MMMU, DocVQA, ChartQA, MathVista, etc. - `2024/02/27`: InternVL is accepted by CVPR 2024 (Oral)! ๐ŸŽ‰ - `2024/02/24`: InternVL-Chat models have been included in the [VLMEvalKit](https://github.com/open-compass/VLMEvalKit). - `2024/02/21`: [InternVL-Chat-V1-2-Plus](https://huggingface.co/OpenGVLab/InternVL-Chat-V1-2-Plus) achieved SOTA performance on MathVista (59.9), MMBench (83.8), and MMVP (58.7). See our [blog](https://internvl.github.io/blog/2024-02-21-InternVL-1.2/) for more details. - `2024/02/12`: InternVL-Chat-V1-2 has been released. It achieves 51.6 on MMMU val and 82.3 on MMBench test. For more details, please refer to our [blog](https://internvl.github.io/blog/2024-02-21-InternVL-1.2/) and [SFT data](./internvl_chat#prepare-training-datasets). The model is now available on [HuggingFace](https://huggingface.co/OpenGVLab/InternVL-Chat-V1-2), and both training / evaluation data and scripts are open-sourced. - `2024/01/24`: InternVL-Chat-V1-1 is released, it supports Chinese and has stronger OCR capability, see [here](https://huggingface.co/OpenGVLab/InternVL-Chat-V1-1). - `2024/01/16`: We release our [customized mmcv/mmsegmentation/mmdetection code](https://github.com/OpenGVLab/InternVL-MMDetSeg), integrated with DeepSpeed, which can be used for training large-scale detection and segmentation models. ## TODO List - [ ] Support vLLM and Ollama - [ ] Rebuild documents using readthedocs - [x] Support fine-tuning different LLMs with LoRA - [ ] Support video and PDF input in online demo - [ ] Release InternVL2 with VisionLLMv2 integration - [x] Release `requirements.txt` for InternVL2 - [x] Release training / evaluation code for InternVL2 series - [x] Release Streamlit web UI for InternVL1.5 and InternVL2 ## Documents - Installation - How to install the environment? [\[link\]](./INSTALLATION.md) [\[requirements.txt\]](./requirements.txt) - Training or Fine-tuning - How to reproduce the SFT stage of InternVL-Chat-V1-2? [\[link\]](./internvl_chat#start-training) - How to fine-tune InternVL-Chat-V1-2 on a custom dataset? [\[link\]](./document/How_to_finetune_internvl_chat_v1_2_on_a_custom_dataset.md) - How to fine-tune the Mini-InternVL-Chat series on a custom dataset? [\[link\]](./document/How_to_finetune_mini_internvl_chat_v1_5_on_a_custom_dataset.md) - Benchmark Test > Due to minor implementation differences between this codebase and VLMEvalKit, slight discrepancies in performance metrics may occur when testing the same model. - How to evaluate InternVL-Chat-V1-5? [\[link\]](./document/How_to_evaluate_internvl_chat_v1_5.md) - How to evaluate InternVL-Chat-V1-5 using VLMEvalKit? (Recommend) [\[link\]](./document/How_to_evaluate_internvl_chat_v1_5_using_vlmevalkit.md) - How to evaluate Mini-InternVL-Chat-2B-V1-5 using VLMEvalKit? (Recommend) [\[link\]](./document/How_to_evaluate_mini_internvl_chat_2b_v1_5_using_vlmevalkit.md) - How to evaluate Mini-InternVL-Chat-4B-V1-5 using VLMEvalKit? (Recommend) [\[link\]](./document/How_to_evaluate_mini_internvl_chat_4b_v1_5_using_vlmevalkit.md) - Deployment - How to use InternVL API? [\[link\]](./document/How_to_use_InternVL_API.md) - How to deploy a local demo? [\[link\]](./document/How_to_deploy_a_local_demo.md) - How to run InternVL-1.5 8bit with Nvidia V100 GPU? [\[link\]](https://github.com/OpenGVLab/InternVL/issues/144) [\[ไธญๆ–‡ๆ•™็จ‹\]](https://zhuanlan.zhihu.com/p/697188143) - How to perform batch inference? [\[link\]](./README.md?plain=1#L849) ## Compared with SOTA VLLMs ![waic_performance](https://github.com/user-attachments/assets/7b24ad6c-45dd-4bcd-aa77-79da1ec856ee) ## Model Zoo #### Multimodal Large Language Model (InternVL 2.0)
Model Name Vision Part Language Part HF Link MS Link Document
InternVL2‑1B InternViT‑300M‑448px Qwen2‑0.5B‑Instruct ๐Ÿค— link ๐Ÿค– link ๐Ÿ“– doc
InternVL2‑2B InternViT‑300M‑448px internlm2‑chat‑1‑8b ๐Ÿค— link ๐Ÿค– link ๐Ÿ“– doc
InternVL2‑4B InternViT‑300M‑448px Phi‑3‑mini‑128k‑instruct ๐Ÿค— link ๐Ÿค– link ๐Ÿ“– doc
InternVL2‑8B InternViT‑300M‑448px internlm2_5‑7b‑chat ๐Ÿค— link ๐Ÿค– link ๐Ÿ“– doc
InternVL2‑26B InternViT‑6B‑448px‑V1‑5 internlm2‑chat‑20b ๐Ÿค— link ๐Ÿค– link ๐Ÿ“– doc
InternVL2‑40B InternViT‑6B‑448px‑V1‑5 Nous‑Hermes‑2‑Yi‑34B ๐Ÿค— link ๐Ÿค– link ๐Ÿ“– doc
InternVL2-Llama3-76B InternViT‑6B‑448px‑V1‑5 Hermesโ€‘2โ€‘Thetaโ€‘
Llamaโ€‘3โ€‘70B		 ๐Ÿค— link ๐Ÿค– link ๐Ÿ“– doc
#### InternVL2-Pro API We encourage everyone to use our API for research. For better management, please submit ([application form](https://docs.google.com/forms/d/e/1FAIpQLSfMCzhPr1OOEKau_6jwTU0EiZMSFckDo-HMlc_hUudhF_97rw/viewform?usp=sf_link)) / ([็”ณ่ฏท่กจ](https://wj.qq.com/s2/14910502/25a4/)) to obtain free API access. #### Multimodal Large Language Model (InternVL 1.0-1.5)
Model Date HF Link MS Link Note
Mini‑InternVL‑Chat‑4B‑V1‑5 2024.05.28 ๐Ÿค— link ๐Ÿค– link ๐Ÿš€๐Ÿš€ 16% of the model size, 90% of the performance
Mini‑InternVL‑Chat‑2B‑V1‑5 2024.05.19 ๐Ÿค— link ๐Ÿค– link ๐Ÿš€ 8% of the model size, 80% of the performance
InternVL‑Chat‑V1‑5 2024.04.18 ๐Ÿค— link ๐Ÿค– link support 4K image; super strong OCR; Approaching the performance of GPT-4V and Gemini Pro on various benchmarks like MMMU, DocVQA, ChartQA, MathVista, etc.
InternVL‑Chat‑V1‑2‑Plus 2024.02.21 ๐Ÿค— link ๐Ÿค– link more SFT data and stronger
InternVL‑Chat‑V1‑2 2024.02.11 ๐Ÿค— link ๐Ÿค– link scaling up LLM to 34B
InternVL‑Chat‑V1‑1 2024.01.24 ๐Ÿค— link ๐Ÿค– link support Chinese and stronger OCR
InternVL‑Chat‑19B 2023.12.25 ๐Ÿค— link ๐Ÿค– link English multimodal dialogue
InternVL‑Chat‑13B 2023.12.25 ๐Ÿค— link ๐Ÿค– link English multimodal dialogue
#### Vision Foundation Model (InternVL 1.0-1.5)
Model Date HF Link MS Link Note
InternViT‑300M‑448px 2024.05.25 ๐Ÿค— link ๐Ÿค– link distilled small vision foundation model with 300M parameters (๐Ÿ”ฅnew)
InternViT‑6B‑448px‑V1‑5 2024.04.20 ๐Ÿค— link ๐Ÿค– link support dynamic resolution and super strong OCR feature extraction capability by incremental pre-training (๐Ÿ”ฅnew)
InternViT‑6B‑448px‑V1‑2 2024.02.11 ๐Ÿค— link ๐Ÿค– link support 448 resolution by incremental pre-training
InternViT‑6B‑448px‑V1‑0 2024.01.30 ๐Ÿค— link ๐Ÿค– link support 448 resolution by incremental pre-training
InternViT‑6B‑224px 2023.12.22 ๐Ÿค— link ๐Ÿค– link the first version of InternViT-6B, extracted from InternVLโ€‘14Bโ€‘224px
#### Vision-Language Foundation Model (InternVL 1.0)
Model Date HF Link MS Link Note
InternVL‑14B‑224px 2023.12.22 ๐Ÿค— link ๐Ÿค– link vision-language foundation model, InternViT-6B + QLLaMA, can be used for image-text retrieval like CLIP
## What can InternVL do?
Visual Perception (click to expand) - Linear-Probe Image Classification [\[see details\]](./classification#-evaluation) ViT-22B uses the private JFT-3B dataset. | method | #param | IN-1K | IN-ReaL | IN-V2 | IN-A | IN-R | IN-Sketch | | ------------------- | :----: | :---: | :-----: | :---: | :--: | :--: | :-------: | | OpenCLIP-G | 1.8B | 86.2 | 89.4 | 77.2 | 63.8 | 87.8 | 66.4 | | DINOv2-g | 1.1B | 86.5 | 89.6 | 78.4 | 75.9 | 78.8 | 62.5 | | EVA-01-CLIP-g | 1.1B | 86.5 | 89.3 | 77.4 | 70.5 | 87.7 | 63.1 | | MAWS-ViT-6.5B | 6.5B | 87.8 | - | - | - | - | - | | ViT-22B\* | 21.7B | 89.5 | 90.9 | 83.2 | 83.8 | 87.4 | - | | InternViT-6B (ours) | 5.9B | 88.2 | 90.4 | 79.9 | 77.5 | 89.8 | 69.1 | - Semantic Segmentation [\[see details\]](./segmentation#-evaluation) | method | decoder | #param (train/total) | crop size | mIoU | | --------------------- | :-----: | :------------------: | :-------: | ------------ | | OpenCLIP-G (frozen) | Linear | 0.3M / 1.8B | 512 | 39.3 | | ViT-22B (frozen) | Linear | 0.9M / 21.7B | 504 | 34.6 | | InternViT-6B (frozen) | Linear | 0.5M / 5.9B | 504 | 47.2 (+12.6) | | ViT-22B (frozen) | UperNet | 0.8B / 22.5B | 504 | 52.7 | | InternViT-6B (frozen) | UperNet | 0.4B / 6.3B | 504 | 54.9 (+2.2) | | ViT-22B | UperNet | 22.5B / 22.5B | 504 | 55.3 | | InternViT-6B | UperNet | 6.3B / 6.3B | 504 | 58.9 (+3.6) | - Zero-Shot Image Classification [\[see details\]](./clip_benchmark#imagenet-variants-and-objectnet) | method | IN-1K | IN-A | IN-R | IN-V2 | IN-Sketch | ObjectNet | | ----------------- | :---: | :--: | :--: | :---: | :-------: | :-------: | | OpenCLIP-G | 80.1 | 69.3 | 92.1 | 73.6 | 68.9 | 73.0 | | EVA-02-CLIP-E+ | 82.0 | 82.1 | 94.5 | 75.7 | 71.6 | 79.6 | | ViT-22B\* | 85.9 | 90.1 | 96.0 | 80.9 | - | 87.6 | | InternVL-C (ours) | 83.2 | 83.8 | 95.5 | 77.3 | 73.9 | 80.6 | - Multilingual Zero-Shot Image Classification [\[see details\]](./clip_benchmark#multilingual-imagenet-1k) EN: English, ZH: Chinese, JP: Japanese, Ar: Arabic, IT: Italian | method | IN-1K (EN) | IN-1K (ZH) | IN-1K (JP) | IN-1K (AR) | IN-1K (IT) | | ----------------- | :--------: | :--------: | :--------: | :--------: | :--------: | | Taiyi-CLIP-ViT-H | - | 54.4 | - | - | - | | WuKong-ViT-L-G | - | 57.5 | - | - | - | | CN-CLIP-ViT-H | - | 59.6 | - | - | - | | AltCLIP-ViT-L | 74.5 | 59.6 | - | - | - | | EVA-02-CLIP-E+ | 82.0 | - | - | - | 41.2 | | OpenCLIP-XLM-R-H | 77.0 | 55.7 | 53.1 | 37.0 | 56.8 | | InternVL-C (ours) | 83.2 | 64.5 | 61.5 | 44.9 | 65.7 | - Zero-Shot Video Classification | method | #frame | K400 | K600 | K700 | | ----------------- | :----: | :--: | :--: | :--: | | OpenCLIP-G | 1 | 65.9 | 66.1 | 59.2 | | EVA-02-CLIP-E+ | 1 | 69.8 | 69.3 | 63.4 | | InternVL-C (ours) | 1 | 71.0 | 71.3 | 65.7 | | ViCLIP | 8 | 75.7 | 73.5 | 66.4 | | InternVL-C (ours) | 8 | 79.4 | 78.8 | 71.5 |
Cross-Modal Retrieval (click to expand) - English Zero-Shot Image-Text Retrieval [\[see details\]](./clip_benchmark#flickr30k--coco)
model Flickr30K COCO avg
image-to-text text-to-image image-to-text text-to-image
R@1 R@5 R@10 R@1 R@5 R@10 R@1 R@5 R@10 R@1 R@5 R@10
OpenCLIP-G 92.9 99.3 99.8 79.5 95.0 97.1 67.3 86.9 92.6 51.4 74.9 83.0 85.0
EVA-02-CLIP-E+ 93.9 99.4 99.8 78.8 94.2 96.8 68.8 87.8 92.8 51.1 75.0 82.7 85.1
EVA-CLIP-8B 95.6 99.6 99.9 80.8 95.5 97.6 70.3 89.3 93.9 53.0 76.0 83.4 86.2
InternVL-C (ours) 94.7 99.6 99.9 81.7 96.0 98.2 70.6 89.0 93.5 54.1 77.3 84.6 86.6
InternVL-G (ours) 95.7 99.7 99.9 85.0 97.0 98.6 74.9 91.3 95.2 58.6 81.3 88.0 88.8
- Chinese Zero-Shot Image-Text Retrieval [\[see details\]](./clip_benchmark#flickr30k-cn--coco-cn)
model Flickr30K-CN COCO-CN avg
image-to-text text-to-image image-to-text text-to-image
R@1 R@5 R@10 R@1 R@5 R@10 R@1 R@5 R@10 R@1 R@5 R@10
CN-CLIP-ViT-H 81.6 97.5 98.8 71.2 91.4 95.5 63.0 86.6 92.9 69.2 89.9 96.1 86.1
OpenCLIP-XLM-R-H 86.1 97.5 99.2 71.0 90.5 94.9 70.0 91.5 97.0 66.1 90.8 96.0 87.6
InternVL-C (ours) 90.3 98.8 99.7 75.1 92.9 96.4 68.8 92.0 96.7 68.9 91.9 96.5 89.0
InternVL-G (ours) 92.9 99.4 99.8 77.7 94.8 97.3 71.4 93.9 97.7 73.8 94.4 98.1 90.9
- Multilingual Zero-Shot Image-Text Retrieval on XTD [\[see details\]](./clip_benchmark#xtd) | method | EN | ES | FR | ZH | IT | KO | RU | JP | average | | ----------------- | :--: | :--: | :--: | :--: | :--: | :--: | :--: | :--: | :-----: | | AltCLIP | 95.4 | 94.1 | 92.9 | 95.1 | 94.2 | 94.4 | 91.8 | 91.7 | 93.7 | | OpenCLIP-XLM-R-H | 97.3 | 96.1 | 94.5 | 94.7 | 96.0 | 90.2 | 93.9 | 94.0 | 94.6 | | InternVL-C (ours) | 97.3 | 95.7 | 95.1 | 95.6 | 96.0 | 92.2 | 93.3 | 95.5 | 95.1 | | InternVL-G (ours) | 98.6 | 97.7 | 96.5 | 96.7 | 96.9 | 95.1 | 94.8 | 96.1 | 96.6 |
Multimodal Dialogue See ["Compared with SOTA VLLMs"](#compared-with-sota-vllms) section.
## Quick Start with HuggingFace
using InternViT-6B for visual feature extraction (click to expand) ```python import torch from PIL import Image from transformers import AutoModel, CLIPImageProcessor model = AutoModel.from_pretrained( 'OpenGVLab/InternViT-6B-448px-V1-5', torch_dtype=torch.bfloat16, low_cpu_mem_usage=True, trust_remote_code=True).cuda().eval() image = Image.open('./examples/image1.jpg').convert('RGB') image_processor = CLIPImageProcessor.from_pretrained('OpenGVLab/InternViT-6B-448px-V1-5') pixel_values = image_processor(images=image, return_tensors='pt').pixel_values pixel_values = pixel_values.to(torch.bfloat16).cuda() outputs = model(pixel_values) ```
using InternVL-C(ontrastive) and InternVL-G(enerative) for cross-modal retrieval (click to expand) ```python import torch from PIL import Image from transformers import AutoModel, CLIPImageProcessor from transformers import AutoTokenizer model = AutoModel.from_pretrained( 'OpenGVLab/InternVL-14B-224px', torch_dtype=torch.bfloat16, low_cpu_mem_usage=True, trust_remote_code=True).cuda().eval() image_processor = CLIPImageProcessor.from_pretrained('OpenGVLab/InternVL-14B-224px') tokenizer = AutoTokenizer.from_pretrained( 'OpenGVLab/InternVL-14B-224px', use_fast=False, add_eos_token=True) tokenizer.pad_token_id = 0 # set pad_token_id to 0 images = [ Image.open('./examples/image1.jpg').convert('RGB'), Image.open('./examples/image2.jpg').convert('RGB'), Image.open('./examples/image3.jpg').convert('RGB') ] prefix = 'summarize:' texts = [ prefix + 'a photo of a red panda', # English prefix + 'ไธ€ๅผ ็†Š็Œซ็š„็…ง็‰‡', # Chinese prefix + 'ไบŒๅŒนใฎ็Œซใฎๅ†™็œŸ' # Japanese ] pixel_values = image_processor(images=images, return_tensors='pt').pixel_values pixel_values = pixel_values.to(torch.bfloat16).cuda() input_ids = tokenizer(texts, return_tensors='pt', max_length=80, truncation=True, padding='max_length').input_ids.cuda() # InternVL-C logits_per_image, logits_per_text = model( image=pixel_values, text=input_ids, mode='InternVL-C') probs = logits_per_image.softmax(dim=-1) # tensor([[9.9609e-01, 5.2185e-03, 6.0070e-08], # [2.2949e-02, 9.7656e-01, 5.9903e-06], # [3.2932e-06, 7.4863e-05, 1.0000e+00]], device='cuda:0', # dtype=torch.bfloat16, grad_fn=) # InternVL-G logits_per_image, logits_per_text = model( image=pixel_values, text=input_ids, mode='InternVL-G') probs = logits_per_image.softmax(dim=-1) # tensor([[9.9609e-01, 3.1738e-03, 3.6322e-08], # [8.6060e-03, 9.9219e-01, 2.8759e-06], # [1.7583e-06, 3.1233e-05, 1.0000e+00]], device='cuda:0', # dtype=torch.bfloat16, grad_fn=) # please set add_eos_token to False for generation tokenizer.add_eos_token = False image = Image.open('./examples/image1.jpg').convert('RGB') pixel_values = image_processor(images=image, return_tensors='pt').pixel_values pixel_values = pixel_values.to(torch.bfloat16).cuda() tokenized = tokenizer("English caption:", return_tensors='pt') pred = model.generate( pixel_values=pixel_values, input_ids=tokenized.input_ids.cuda(), attention_mask=tokenized.attention_mask.cuda(), num_beams=5, min_new_tokens=8, ) caption = tokenizer.decode(pred[0].cpu(), skip_special_tokens=True).strip() # English caption: a red panda sitting on top of a wooden platform ```
using InternVL-Chat for multimodal chat (click to expand) Here, we take the smaller OpenGVLab/InternVL2-8B as an example: ```python import numpy as np import torch import torchvision.transforms as T from decord import VideoReader, cpu from PIL import Image from torchvision.transforms.functional import InterpolationMode from transformers import AutoModel, AutoTokenizer IMAGENET_MEAN = (0.485, 0.456, 0.406) IMAGENET_STD = (0.229, 0.224, 0.225) def build_transform(input_size): MEAN, STD = IMAGENET_MEAN, IMAGENET_STD transform = T.Compose([ T.Lambda(lambda img: img.convert('RGB') if img.mode != 'RGB' else img), T.Resize((input_size, input_size), interpolation=InterpolationMode.BICUBIC), T.ToTensor(), T.Normalize(mean=MEAN, std=STD) ]) return transform def find_closest_aspect_ratio(aspect_ratio, target_ratios, width, height, image_size): best_ratio_diff = float('inf') best_ratio = (1, 1) area = width * height for ratio in target_ratios: target_aspect_ratio = ratio[0] / ratio[1] ratio_diff = abs(aspect_ratio - target_aspect_ratio) if ratio_diff < best_ratio_diff: best_ratio_diff = ratio_diff best_ratio = ratio elif ratio_diff == best_ratio_diff: if area > 0.5 * image_size * image_size * ratio[0] * ratio[1]: best_ratio = ratio return best_ratio def dynamic_preprocess(image, min_num=1, max_num=6, image_size=448, use_thumbnail=False): orig_width, orig_height = image.size aspect_ratio = orig_width / orig_height # calculate the existing image aspect ratio target_ratios = set( (i, j) for n in range(min_num, max_num + 1) for i in range(1, n + 1) for j in range(1, n + 1) if i * j <= max_num and i * j >= min_num) target_ratios = sorted(target_ratios, key=lambda x: x[0] * x[1]) # find the closest aspect ratio to the target target_aspect_ratio = find_closest_aspect_ratio( aspect_ratio, target_ratios, orig_width, orig_height, image_size) # calculate the target width and height target_width = image_size * target_aspect_ratio[0] target_height = image_size * target_aspect_ratio[1] blocks = target_aspect_ratio[0] * target_aspect_ratio[1] # resize the image resized_img = image.resize((target_width, target_height)) processed_images = [] for i in range(blocks): box = ( (i % (target_width // image_size)) * image_size, (i // (target_width // image_size)) * image_size, ((i % (target_width // image_size)) + 1) * image_size, ((i // (target_width // image_size)) + 1) * image_size ) # split the image split_img = resized_img.crop(box) processed_images.append(split_img) assert len(processed_images) == blocks if use_thumbnail and len(processed_images) != 1: thumbnail_img = image.resize((image_size, image_size)) processed_images.append(thumbnail_img) return processed_images def load_image(image_file, input_size=448, max_num=6): image = Image.open(image_file).convert('RGB') transform = build_transform(input_size=input_size) images = dynamic_preprocess(image, image_size=input_size, use_thumbnail=True, max_num=max_num) pixel_values = [transform(image) for image in images] pixel_values = torch.stack(pixel_values) return pixel_values path = 'OpenGVLab/InternVL2-8B' model = AutoModel.from_pretrained( path, torch_dtype=torch.bfloat16, low_cpu_mem_usage=True, trust_remote_code=True).eval().cuda() tokenizer = AutoTokenizer.from_pretrained(path, trust_remote_code=True) # set the max number of tiles in `max_num` pixel_values = load_image('./examples/image1.jpg', max_num=6).to(torch.bfloat16).cuda() generation_config = dict( num_beams=1, max_new_tokens=1024, do_sample=False, ) # pure-text conversation (็บฏๆ–‡ๆœฌๅฏน่ฏ) question = 'Hello, who are you?' response, history = model.chat(tokenizer, None, question, generation_config, history=None, return_history=True) print(f'User: {question}') print(f'Assistant: {response}') question = 'Can you tell me a story?' response, history = model.chat(tokenizer, None, question, generation_config, history=history, return_history=True) print(f'User: {question}') print(f'Assistant: {response}') # single-image single-round conversation (ๅ•ๅ›พๅ•่ฝฎๅฏน่ฏ) question = '\nPlease describe the image shortly.' response = model.chat(tokenizer, pixel_values, question, generation_config) print(f'User: {question}') print(f'Assistant: {response}') # single-image multi-round conversation (ๅ•ๅ›พๅคš่ฝฎๅฏน่ฏ) question = '\nPlease describe the image in detail.' response, history = model.chat(tokenizer, pixel_values, question, generation_config, history=None, return_history=True) print(f'User: {question}') print(f'Assistant: {response}') question = 'Please write a poem according to the image.' response, history = model.chat(tokenizer, pixel_values, question, generation_config, history=history, return_history=True) print(f'User: {question}') print(f'Assistant: {response}') # multi-image multi-round conversation, combined images (ๅคšๅ›พๅคš่ฝฎๅฏน่ฏ๏ผŒๆ‹ผๆŽฅๅ›พๅƒ) pixel_values1 = load_image('./examples/image1.jpg', max_num=6).to(torch.bfloat16).cuda() pixel_values2 = load_image('./examples/image2.jpg', max_num=6).to(torch.bfloat16).cuda() pixel_values = torch.cat((pixel_values1, pixel_values2), dim=0) question = '\nDescribe the two images in detail.' response, history = model.chat(tokenizer, pixel_values, question, generation_config, history=None, return_history=True) question = 'What are the similarities and differences between these two images.' response, history = model.chat(tokenizer, pixel_values, question, generation_config, history=history, return_history=True) print(f'User: {question}') print(f'Assistant: {response}') # multi-image multi-round conversation, separate images (ๅคšๅ›พๅคš่ฝฎๅฏน่ฏ๏ผŒ็‹ฌ็ซ‹ๅ›พๅƒ) pixel_values1 = load_image('./examples/image1.jpg', max_num=6).to(torch.bfloat16).cuda() pixel_values2 = load_image('./examples/image2.jpg', max_num=6).to(torch.bfloat16).cuda() pixel_values = torch.cat((pixel_values1, pixel_values2), dim=0) num_patches_list = [pixel_values1.size(0), pixel_values2.size(0)] question = 'Image-1: \nImage-2: \nDescribe the two images in detail.' response, history = model.chat(tokenizer, pixel_values, question, generation_config, num_patches_list=num_patches_list, history=None, return_history=True) print(f'User: {question}') print(f'Assistant: {response}') question = 'What are the similarities and differences between these two images.' response, history = model.chat(tokenizer, pixel_values, question, generation_config, num_patches_list=num_patches_list, history=history, return_history=True) print(f'User: {question}') print(f'Assistant: {response}') # batch inference, single image per sample (ๅ•ๅ›พๆ‰นๅค„็†) pixel_values1 = load_image('./examples/image1.jpg', max_num=6).to(torch.bfloat16).cuda() pixel_values2 = load_image('./examples/image2.jpg', max_num=6).to(torch.bfloat16).cuda() num_patches_list = [pixel_values1.size(0), pixel_values2.size(0)] pixel_values = torch.cat((pixel_values1, pixel_values2), dim=0) questions = ['\nDescribe the image in detail.'] * len(num_patches_list) responses = model.batch_chat(tokenizer, pixel_values, num_patches_list=num_patches_list, questions=questions, generation_config=generation_config) for question, response in zip(questions, responses): print(f'User: {question}') print(f'Assistant: {response}') # video multi-round conversation (่ง†้ข‘ๅคš่ฝฎๅฏน่ฏ) def get_index(bound, fps, max_frame, first_idx=0, num_segments=32): if bound: start, end = bound[0], bound[1] else: start, end = -100000, 100000 start_idx = max(first_idx, round(start * fps)) end_idx = min(round(end * fps), max_frame) seg_size = float(end_idx - start_idx) / num_segments frame_indices = np.array([ int(start_idx + (seg_size / 2) + np.round(seg_size * idx)) for idx in range(num_segments) ]) return frame_indices def load_video(video_path, bound=None, input_size=448, max_num=1, num_segments=32): vr = VideoReader(video_path, ctx=cpu(0), num_threads=1) max_frame = len(vr) - 1 fps = float(vr.get_avg_fps()) pixel_values_list, num_patches_list = [], [] transform = build_transform(input_size=input_size) frame_indices = get_index(bound, fps, max_frame, first_idx=0, num_segments=num_segments) for frame_index in frame_indices: img = Image.fromarray(vr[frame_index].asnumpy()).convert('RGB') img = dynamic_preprocess(img, image_size=input_size, use_thumbnail=True, max_num=max_num) pixel_values = [transform(tile) for tile in img] pixel_values = torch.stack(pixel_values) num_patches_list.append(pixel_values.shape[0]) pixel_values_list.append(pixel_values) pixel_values = torch.cat(pixel_values_list) return pixel_values, num_patches_list video_path = './examples/red-panda.mp4' # pixel_values, num_patches_list = load_video(video_path, num_segments=32, max_num=1) pixel_values, num_patches_list = load_video(video_path, num_segments=8, max_num=1) pixel_values = pixel_values.to(torch.bfloat16).cuda() video_prefix = ''.join([f'Frame{i+1}: \n' for i in range(len(num_patches_list))]) question = video_prefix + 'What is the red panda doing?' # Frame1: \nFrame2: \n...\nFrame31: \n{question} response, history = model.chat(tokenizer, pixel_values, question, generation_config, num_patches_list=num_patches_list, history=None, return_history=True) print(f'User: {question}') print(f'Assistant: {response}') question = 'Describe this video in detail. Don\'t repeat.' response, history = model.chat(tokenizer, pixel_values, question, generation_config, num_patches_list=num_patches_list, history=history, return_history=True) print(f'User: {question}') print(f'Assistant: {response}') ```
## License This project is released under the [MIT license](LICENSE). Parts of this project contain code and models from other sources, which are subject to their respective licenses. ## Citation If you find this project useful in your research, please consider cite: ```BibTeX @article{chen2023internvl, title={InternVL: Scaling up Vision Foundation Models and Aligning for Generic Visual-Linguistic Tasks}, author={Chen, Zhe and Wu, Jiannan and Wang, Wenhai and Su, Weijie and Chen, Guo and Xing, Sen and Zhong, Muyan and Zhang, Qinglong and Zhu, Xizhou and Lu, Lewei and Li, Bin and Luo, Ping and Lu, Tong and Qiao, Yu and Dai, Jifeng}, journal={arXiv preprint arXiv:2312.14238}, year={2023} } @article{chen2024far, title={How Far Are We to GPT-4V? Closing the Gap to Commercial Multimodal Models with Open-Source Suites}, author={Chen, Zhe and Wang, Weiyun and Tian, Hao and Ye, Shenglong and Gao, Zhangwei and Cui, Erfei and Tong, Wenwen and Hu, Kongzhi and Luo, Jiapeng and Ma, Zheng and others}, journal={arXiv preprint arXiv:2404.16821}, year={2024} } ``` ## Acknowledgement InternVL is built with reference to the code of the following projects: [OpenAI CLIP](https://github.com/openai/CLIP), [Open CLIP](https://github.com/mlfoundations/open_clip), [CLIP Benchmark](https://github.com/LAION-AI/CLIP_benchmark), [EVA](https://github.com/baaivision/EVA/tree/master), [InternImage](https://github.com/OpenGVLab/InternImage), [ViT-Adapter](https://github.com/czczup/ViT-Adapter), [MMSegmentation](https://github.com/open-mmlab/mmsegmentation), [Transformers](https://github.com/huggingface/transformers), [DINOv2](https://github.com/facebookresearch/dinov2), [BLIP-2](https://github.com/salesforce/LAVIS/tree/main/projects/blip2), [Qwen-VL](https://github.com/QwenLM/Qwen-VL/tree/master/eval_mm), and [LLaVA-1.5](https://github.com/haotian-liu/LLaVA). Thanks for their awesome work! ______________________________________________________________________ If you want to join our WeChat group, please scan the following QR Code to add our assistant as a Wechat friend:

image