Image Feature Extraction
OpenCLIP
English
clip
open-clip
siglip
fashion
text-to-image-retrieval
multimodal-retrieval
embedding
mps
Instructions to use HopitAI/moda-fashionsiglip-multiview-203m with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- OpenCLIP
How to use HopitAI/moda-fashionsiglip-multiview-203m with OpenCLIP:
import open_clip model, preprocess_train, preprocess_val = open_clip.create_model_and_transforms('hf-hub:HopitAI/moda-fashionsiglip-multiview-203m') tokenizer = open_clip.get_tokenizer('hf-hub:HopitAI/moda-fashionsiglip-multiview-203m') - Notebooks
- Google Colab
- Kaggle
| { | |
| "schema_version": 1, | |
| "artifact_kind": "public_fashionsiglip_late_fusion_summary", | |
| "status": "AT_LEAST_4_OF_6_ACHIEVED", | |
| "frozen_at_date": "2026-07-29", | |
| "candidate": { | |
| "candidate_id": "late-maxview-b010", | |
| "base_model": "Marqo/marqo-fashionSigLIP", | |
| "neural_parameters": 203155970, | |
| "additional_learned_parameters": 0, | |
| "embedding_dimension": 768, | |
| "stored_document_vectors": 3, | |
| "ann_routes": 3, | |
| "formula": "0.9*parent+0.1*max(parent,pad,center_crop)" | |
| }, | |
| "protocol": { | |
| "task": "text-to-image", | |
| "full_corpus": true, | |
| "primary_metric": "MAP@10", | |
| "paired_bootstrap_samples": 10000, | |
| "same_global_recipe_all_datasets": true, | |
| "fresh_blind_sota_evaluation": false, | |
| "benchmark_iteration": 6, | |
| "target_examples_images_qrels_used_for_selection": false | |
| }, | |
| "datasets": { | |
| "KAGL": { | |
| "candidate_MAP@10": 0.29074, | |
| "baseline_MAP@10": 0.27687, | |
| "absolute_delta": 0.013870606533017249, | |
| "ci95_low": 0.008940084980330517, | |
| "ci95_high": 0.0188327481806801, | |
| "classification": "significant_win" | |
| }, | |
| "fashion200k": { | |
| "candidate_MAP@10": 0.1951, | |
| "baseline_MAP@10": 0.18577, | |
| "absolute_delta": 0.009324404761904763, | |
| "ci95_low": 0.003819347718253968, | |
| "ci95_high": 0.014771639384920635, | |
| "classification": "significant_win" | |
| }, | |
| "deepfashion_inshop": { | |
| "candidate_MAP@10": 0.16371, | |
| "baseline_MAP@10": 0.15865, | |
| "absolute_delta": 0.005066924557275024, | |
| "ci95_low": 0.0030408831185772256, | |
| "ci95_high": 0.007095991747488355, | |
| "classification": "significant_win" | |
| }, | |
| "polyvore": { | |
| "candidate_MAP@10": 0.37191, | |
| "baseline_MAP@10": 0.36645, | |
| "absolute_delta": 0.005465674603174603, | |
| "ci95_low": 0.0008606845238095241, | |
| "ci95_high": 0.01014109623015873, | |
| "classification": "significant_win" | |
| }, | |
| "atlas": { | |
| "candidate_MAP@10": 0.18637, | |
| "baseline_MAP@10": 0.18264, | |
| "absolute_delta": 0.003737096945082783, | |
| "ci95_low": -0.00016614619280408522, | |
| "ci95_high": 0.007746641375037741, | |
| "classification": "inconclusive" | |
| }, | |
| "deepfashion_multimodal": { | |
| "candidate_MAP@10": 0.01504, | |
| "baseline_MAP@10": 0.01477, | |
| "absolute_delta": 0.0002765873015873017, | |
| "ci95_low": -0.0012055704365079366, | |
| "ci95_high": 0.001916284722222222, | |
| "classification": "inconclusive" | |
| } | |
| }, | |
| "evaluated_datasets": 6, | |
| "significant_wins": 4, | |
| "significant_losses": 0, | |
| "point_wins": 6 | |
| } | |