TinyCLIP ViT-39M/16 (Core ML) for Lossless Video Cutter
The model Lossless Video Cutter downloads for Find by Description's standard mode. It is
Microsoft's TinyCLIP ViT-39M/16 Text-19M (YFCC15M)
(commit 07a4b0bc751cb64fecd2b661c048c1dd98d69444), converted to Core ML by
macos/scripts/convert_tinyclip.py:
Image.mlpackage: 224 × 224 RGB in, L2-normalized 512-value vector out.Text.mlpackage: 77 CLIP token ids in, L2-normalized 512-value vector out.tokenizer-vocab.json,tokenizer-merges.txt: the CLIP BPE tokenizer.provenance.json: source commit and hashes of the conversion.
Licensed under the MIT License, like the original model; see NOTICE.
- Downloads last month
- 40
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support
Model tree for nufrnd/lvc-tinyclip-coreml
Base model
wkcn/TinyCLIP-ViT-39M-16-Text-19M-YFCC15M