| 123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899 |
- Collections:
- - Name: BLIP
- Metadata:
- Training Data:
- - COCO
- - VG
- - Conceptual Captions
- - Conceptual 12M
- - SBU captions
- Architecture:
- - Transformer
- Training Resources: 8x A100 GPUs
- Paper:
- Title: 'BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language
- Understanding and Generation'
- URL: https://arxiv.org/abs/2201.12086
- README: configs/blip/README.md
- Models:
- - Name: blip-base_8xb16_refcoco
- Metadata:
- FLOPs: null
- Parameters: 498488636
- In Collection: BLIP
- Results:
- - Task: Visual Grounding
- Dataset: RefCOCO
- Metrics:
- Accuracy (testA): 86.14
- Accuracy (testB): 77.33
- Weights: https://download.openmmlab.com/mmclassification/v1/blip/blip-base_8xb16_refcoco_20230508-d2d10f4c.pth
- Config: configs/blip/blip-base_8xb16_refcoco.py
- - Name: blip-base_3rdparty_caption
- Metadata:
- FLOPs: null
- Parameters: 223971644
- In Collection: BLIP
- Results:
- - Dataset: COCO
- Task: Image Caption
- Metrics:
- BLEU-4: 40.12
- CIDER: 132.82
- Weights: https://download.openmmlab.com/mmclassification/v1/blip/blip-base_3rdparty_coco-caption_20230419-a5b71af3.pth
- Config: configs/blip/blip-base_8xb32_caption.py
- Converted From:
- Weights: https://storage.googleapis.com/sfr-vision-language-research/LAVIS/models/BLIP/blip_coco_caption_base.pth
- Code: https://github.com/salesforce/LAVIS
- - Name: blip-base_3rdparty_nlvr
- Metadata:
- FLOPs: null
- Parameters: 259372034
- In Collection: BLIP
- Results:
- - Task: NLVR
- Dataset: NLVR2
- Metrics:
- Top 1 Accuracy: 82.33
- Weights: https://download.openmmlab.com/mmclassification/v1/blip/blip-base_3rdparty_nlvr_20230427-3b14d33f.pth
- Config: configs/blip/blip-base_8xb32_nlvr.py
- Converted From:
- Weights: https://storage.googleapis.com/sfr-vision-language-research/BLIP/models/model_base_nlvr.pth
- Code: https://github.com/salesforce/LAVIS
- - Name: blip-base_3rdparty_vqa
- Metadata:
- FLOPs: null
- Parameters: 361478972
- In Collection: BLIP
- Results:
- - Task: Visual Question Answering
- Dataset: VQAv2
- Metrics:
- Accuracy: 78.2
- Weights: https://download.openmmlab.com/mmclassification/v1/blip/blip-base_3rdparty-capflit_vqa_20230505-81488941.pth
- Config: configs/blip/blip-base_8xb32_vqa.py
- Converted From:
- Weights: https://storage.googleapis.com/sfr-vision-language-research/BLIP/models/model_base_vqa_capfilt_large.pth
- Code: https://github.com/salesforce/LAVIS
- - Name: blip-base_3rdparty_retrieval
- Metadata:
- FLOPs: null
- Parameters: 447486979
- In Collection: BLIP
- Results:
- - Task: Image-To-Text Retrieval
- Dataset: COCO
- Metrics:
- Recall@1: 82.52
- Recall@5: 95.34
- - Task: Text-To-Image Retrieval
- Dataset: COCO
- Metrics:
- Recall@1: 64.82
- Recall@5: 86.28
- Weights: https://download.openmmlab.com/mmclassification/v1/blip/blip-base_3rdparty_coco-retrieval_20230419-a1804d2c.pth
- Config: configs/blip/blip-base_8xb32_retrieval.py
- Converted From:
- Weights: https://storage.googleapis.com/sfr-vision-language-research/LAVIS/models/BLIP/blip_coco_retrieval.pth
- Code: https://github.com/salesforce/LAVIS
|
