Model reference · open weights
mlcd-vit-bigG-patch14-448 is an open-weight embedding model from DeepGlint-AI. mlcd-vit-bigG-patch14-448 (FP32) weighs 3.7 GB; the smallest configuration that runs it is RTX 3060 12 GB.
What it is
| Released by | DeepGlint-AI |
|---|---|
| Type | Embedding models |
| Task | Image embed |
| Parameters (lead) | 1.8B |
| Released | 2025-02-12 |
| Popularity | 667 downloads / month |
| Weights | 3.7 GB (mlcd-vit-bigG-patch14-448 (FP32), file size) |
| Licence | Open weights |
What it runs on
Weights 3.7 GB (file size) · overhead about 1.1 GB.
| Card | Runs | Counted memory |
|---|---|---|
| RTX 3060 12 GB | fits | 11.6 GB |
| RTX 4060 Ti 16 GB | fits | 15.4 GB |
| RTX 3090 24 GB | fits | 23.4 GB |
| RTX 4090 24 GB | fits | 23.4 GB |
| RTX 5090 32 GB | fits | 31.0 GB |
| L40S 48 GB | fits | 44.0 GB |
| A100 80 GB | fits | 78.2 GB |
| H100 80 GB | fits | 78.1 GB |
| RTX PRO 6000 Blackwell 96 GB | fits | 93.8 GB |
| DGX Spark (GB10) 128 GB unified | fits | 107 GB |
| H200 141 GB | fits | 138 GB |
| B200 180 GB | fits | 176 GB |
Estimates, not measurements: the weights are the build's file size. No cache grows with use; a batch of inputs needs working memory of its own. Counted memory is 92 % of what CUDA reports for the card.
From the model card
[!TIP] LLaVA-NeXT and transformers now supports MLCD-ViT-bigG-14-448px.
MLCD-ViT-bigG is a state-of-the-art vision transformer model enhanced with 2D Rotary Position Embedding (RoPE2D), achieving superior performance on document understanding and visual question answering tasks. Developed by DeepGlint AI, this model demonstrates exceptional capabilities in processing complex visual-language interactions.
We adopted the official LLaVA-NeXT and the official training dataset LLaVA-NeXT-Data for evaluating the foundational visual models. The language model is Qwen2.5-7B.
| Vision Tower | RoPE2D | ChartQA | DocVQA | InfoVQA | OCRBench | MMMU |
|---|---|---|---|---|---|---|
| CLIP (ViT-L-14-336px) | × | 66.52 | 75.21 | 38.88 | 525.00 | 44.20 |
| SigLIP (ViT-SO400M-384px) | × | 69.28 | 76.71 | 41.38 | 554.00 | 46.78 |
| DFN5B (ViT-H-14-378px) | × | 64.36 | 70.87 | 38.59 | 473.00 | 48.00 |
| MLCD (ViT-L-14-336px) | × | 67.84 | 76.46 | 43.48 | 531.00 | 44.30 |
| MLCD (ViT-bigG-14-336px) | √ | 71.07 | 79.63 | 44.38 | 572.00 | 46.78 |
| MLCD (ViT-bigG-14-448px) | √ | 73.80 | 83.34 | 46.59 | 582.00 | 46.00 |
pip install torch transformers
git clone https://github.com/deepglint/unicom
cd unicom/mlcd
from vit_rope2d_hf import MLCDVisionModel
from transformers import CLIPImageProcessor
from PIL import Image
import requests
import torch
# Load model and processor
model = MLCDVisionModel.from_pretrained("DeepGlint-AI/mlcd-vit-bigG-patch14-448")
processor = CLIPImageProcessor.from_pretrained("DeepGlint-AI/mlcd-vit-bigG-patch14-448")
# Process single image
url = "http://images.cocodataset.org/val2017/000000039769.jpg"
image = Image.open(requests.get(url, stream=True).raw)
inputs = processor(images=image, return_tensors="pt")
# Get visual features
with torch.no_grad():
outputs = model(**inputs)
features = outputs.last_hidden_state
print(f"Extracted features shape: {features.shape}")
@inproceedings{anxiang_2024_mlcd,
title={Multi-label Cluster Discrimination for Visual Representation Learning},
author={An, Xiang and Yang, Kaicheng and Dai, Xiangzi and Feng, Ziyong and Deng, Jiankang},
booktitle={ECCV},
year={2024}
}
Quoted from the model card on Hugging Face — the full card is behind the Hugging Face link above.