Upload README.md with huggingface_hub

README.md

@@ -1,6 +1,6 @@
 ---
 license: cc-by-nc-4.0
-pretty_name: HTT dataset
+pretty_name: HTT Dataset & Model
 tags:
 - robotics
 - tactile-sensing
@@ -9,11 +9,12 @@ size_categories:
 - 1M<n<10M
 ---
 
-# HTT dataset
+# HTT Dataset & Model
 
-Paired multi-sensor tactile data (**1.59M frames**) for the Heterogeneous
-Tactile Transformer (HTT). Four sensors — two vision-based, two taxel arrays —
-organized into four task splits.
+Paired multi-sensor tactile data (**1.59M frames**) **and the pretrained
+model checkpoint** for the Heterogeneous Tactile Transformer (HTT). Four
+sensors — two vision-based, two taxel arrays — organized into four task
+splits.
 
 | sensor | type | raw format |
 |---|---|---|
@@ -36,15 +37,16 @@ organized into four task splits.
 
 ## Model checkpoint
 
-The pretrained HTT model (`htt_4sensors_best.pth`, ~69 MB, SHA-256
-`024f4c3a067168197d0a6996bbca5c03e744ed5abd1d35a666dbf78e7ac673f0`) is bundled
-here for convenience and also hosted at
-[AllenBi21/HTT](https://huggingface.co/AllenBi21/HTT):
+This repo is the single home for **dataset & model**. The pretrained HTT
+checkpoint (`htt_4sensors_best.pth`, ~69 MB, SHA-256
+`024f4c3a067168197d0a6996bbca5c03e744ed5abd1d35a666dbf78e7ac673f0`):
 
 ```bash
 hf download AllenBi21/HTT-dataset htt_4sensors_best.pth --repo-type dataset --local-dir checkpoints
 ```
 
+Inference API and training code: [github.com/jxbi1010/HTT](https://github.com/jxbi1010/HTT).
+
 ## Statistics
 
 One *paired sample* = 1 optical frame + 1 synchronized chunk of 10 taxel