Visual Question Answering
Transformers
Safetensors
baby_medgemma
feature-extraction
medical
chest-x-ray
vision-language-model
interpretability
mechanistic-interpretability
research
custom_code
Instructions to use saillab/FlipLens with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use saillab/FlipLens with Transformers:
# Use a pipeline as a high-level helper # Warning: Pipeline type "visual-question-answering" is no longer supported in transformers v5. # You must load the model directly (see below) or downgrade to v4.x with: # pip install "transformers<5.0.0" from transformers import pipeline pipe = pipeline("visual-question-answering", model="saillab/FlipLens", trust_remote_code=True)# pip install -U transformers accelerate # Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("saillab/FlipLens", trust_remote_code=True, device_map="auto") - Notebooks
- Google Colab
- Kaggle
Commit History
-
Rename model card to FlipLens; fix the text-only range to within one point of chance
Verified signature2819a50 -
Correct shuffled-grounding AUC to 0.502 (was 0.506)
Verified signature20ee8e4
-
Add the single-pass margin-gate flip detector and the route-flip null result
Verified signature5efcebc
-
Card: single model
Verified signature4136cc2 -
Keep only the latest model: remove the retired 1,841-question probe and its artifacts
Verified signature9c2651d -
Card: document repository structure and link the code on GitHub
Verified signature90996c8
-
Card: provenance at scale, held-out phrasing, threshold robustness
Verified signature9356d18 -
Card: document both variants and transfer results
Verified signatureb96a912 -
Scaled grounded model: 107k balanced questions, transfers to unseen hospitals
Verified signature19f15e4 -
Preserve the 1,841-question dissertation probe under probe-1841/
Verified signaturee23576b
-
Refresh D checkpoints (896 + LM-head)
Verified signature5b589dd -
Refresh C checkpoints (896 + LM-head)
Verified signatureb77bfa7 -
Refresh B checkpoints (896 + LM-head)
Verified signature7c52f5e -
Feature cache recomputed at 896 pixels
Verified signatureeaba0ff -
Bundled probe: 896 resolution + tied LM-head readout
Verified signature3e03dd6 -
Add model card (896 + LM-head faithful release)
Verified signature712de7a
-
Card: point experiments link to the repo README (EXPERIMENTS.md merged in)
Verified signature54e9276
-
Card: explicit scope note + link to what the experiments prove
Verified signature05e4e2c -
Model card: transformers load instructions
Verified signature06b4d96 -
Add transformers-loadable model (config + safetensors + modeling code)
Verified signature7afb903 -
Add proper model card with metadata
Verified signature356945a -
Upload folder using huggingface_hub
Verified signaturecf48f7a -
Upload feature_cache/medsiglip_feats.pt with huggingface_hub
Verified signaturea61a558 -
Upload README.md with huggingface_hub
Verified signaturec234d30 -
Upload README.md with huggingface_hub
Verified signature39be930 -
initial commit
Verified signature2388037