Card: positioning alt text matches the trained-parameters axis b82ccac verified abtonmoy commited on about 17 hours ago
Positioning figure: trained-parameters axis (matches the technical report) 415eb5c verified abtonmoy commited on about 17 hours ago
Remote code: batched embed_text_batch / embed_audio_batch entry points e6d91bc verified abtonmoy commited on about 19 hours ago
AutoModel + trust_remote_code support (remote-code loading; existing loaders unchanged) 27e1dd7 verified abtonmoy commited on about 19 hours ago
MAEB: exact FSDKaggle2019 overlap numbers (608/4481 verified by Freesound id) 87f90d8 verified abtonmoy commited on 4 days ago
MAEB: FSD2019Kaggle disclosure (anonymized filenames, disjointness unverifiable) a65af61 verified abtonmoy commited on 4 days ago
Restructure card: architecture up top, unified-model comparison leads evaluation 5020012 verified abtonmoy commited on 4 days ago
Add Gemini Embedding 2 baseline to cross-modal comparison 49f04ad verified abtonmoy commited on 4 days ago
v0.3: AudioCaps in-domain fine-tune stage (a2t R@10 0.741, R@1 0.332) e92d9df verified abtonmoy commited on 5 days ago
Model card: architecture diagram; fix bank size to 484K (v0.2) bf1eb1a verified abtonmoy commited on 6 days ago
v0.2: 484K-pair corpus checkpoint (AudioCaps 0.717, Clotho-ZS 0.448) 4ba4d39 verified abtonmoy commited on 6 days ago
Add full cross-modal metrics and qualitative retrieval examples 5874cec verified abtonmoy commited on 6 days ago
Upload fusion-embedding-1-2b-preview.pt with huggingface_hub 170cc2e verified abtonmoy commited on 7 days ago