egrace479 commited on
Commit
d09d6f8
1 Parent(s): 454d409

Update README.md

Browse files

Update following Rare Species data adjustment, consistent with Rare Species dataset as published (https://doi.org/10.57967/hf/1511).

Files changed (1) hide show
  1. README.md +1 -1
README.md CHANGED
@@ -33,7 +33,7 @@ This modelcard has been generated using [this raw template](https://github.com/h
33
 
34
  BioCLIP is a foundation model for the tree of life, built using CLIP architecture as a vision model for general organismal biology.
35
  It is trained on [TreeOfLife-10M](https://huggingface.co/datasets/imageomics/TreeOfLife-10M), our specially-created dataset covering over 450K taxa--the most biologically diverse ML-ready dataset available to date.
36
- Through rigorous benchmarking on a diverse set of fine-grained biological classification tasks, BioCLIP consistently outperformed existing baselines by 17% to 20% absolute.
37
  Through intrinsic evaluation, we found that BioCLIP learned a hierarchical representation aligned to the tree of life, which demonstrates its potential for robust generalizability.
38
 
39
  **See the `examples/` directory for examples of how to use BioCLIP in zero-shot and few-shot settings.**
 
33
 
34
  BioCLIP is a foundation model for the tree of life, built using CLIP architecture as a vision model for general organismal biology.
35
  It is trained on [TreeOfLife-10M](https://huggingface.co/datasets/imageomics/TreeOfLife-10M), our specially-created dataset covering over 450K taxa--the most biologically diverse ML-ready dataset available to date.
36
+ Through rigorous benchmarking on a diverse set of fine-grained biological classification tasks, BioCLIP consistently outperformed existing baselines by 16% to 17% absolute.
37
  Through intrinsic evaluation, we found that BioCLIP learned a hierarchical representation aligned to the tree of life, which demonstrates its potential for robust generalizability.
38
 
39
  **See the `examples/` directory for examples of how to use BioCLIP in zero-shot and few-shot settings.**