Update README.md

README.md

@@ -11,11 +11,11 @@ tags:
 - pathology
 license: apache-2.0
 ---
-# cx-ts:tcga-v1
+# CxTissueSeg
 
 ## Overview
 
-The **cx-ts:tcga-v1** model performs binary segmentation of patches of tissue present in [H&E](https://en.wikipedia.org/wiki/H%26E_stain) pathology slides.
+The **CxTissueSeg** model performs binary segmentation of patches of tissue present in [H&E](https://en.wikipedia.org/wiki/H%26E_stain) pathology slides.
 It is architected to run efficiently on resource constrained systems, providing tissue segmentation on a slide in under 1 second on a typical CPU.
 
 The model is trained on a manually curated set of slides from [our linked dataset](https://huggingface.co/datasets/conflux-xyz/tcga-tissue-segmentation), where it achieves 0.93 mIoU for tissue on the test split.
@@ -36,7 +36,7 @@ For more details on the background of the model, check out the blog post here: h
 
 ## Usage
 
-**cx-ts:tgca-v1** was trained on 512 x 512 pixel patches from thumbnail images of whole slides at 40 microns per pixel (MPP) -- a 4x downsample from the images in the dataset.
+**CxTissueSeg** was trained on 512 x 512 pixel patches from thumbnail images of whole slides at 40 microns per pixel (MPP) -- a 4x downsample from the images in the dataset.
 Thus, it is important when running inference with the model to run it on 40 MPP thumbnails and run inference on tiles of the same dimension (512 x 512).
 When padding tiles, pad with pure white: `rgb(255, 255, 255)`.