openai/clip-vit-large-patch14
Primitive: /encode · Encode ·
CLIP
Disclaimer: The model card is taken and modified from the official CLIP repository, it can be found here.
MultimodalDense
View on Hugging Face → Guide: CLIP family guide →
Overview
Hardware: — drives latency, throughput & cost
| Size | 428M params |
|---|---|
| Tasks | /encode |
| License | — |
| Latency | 228 ms |
| Throughput | 977 tok/s |
| Cost | $0.227 /1M tok |
Cost is approximate — computed from list GPU prices; your actual price depends on the provider you deploy SIE with.
Embedding
| Output types | Dense |
|---|---|
| Dimensions | dense: 768 |
| Max sequence length | 77 |
| Inputs | text · image |
Benchmarks
Flickr30kI2TRetrieval
Image-to-text retrieval: retrieve captions from images
Corpus: 31,783 Queries: 1,000
Quality
ndcg at 10 0.7824
map at 10 0.6816
mrr at 10 0.9111
Performance L4 b1 c16
Corpus 977 tok/s
Corpus p50 228.0ms
Query 7.6 mpix/s
Query p50 340.1ms
Compare (0)Compare →