Update README.md

- **1,105** lifestyle videos that span diverse scenarios (from ActivityNet).
- The videos can be found in ActivityNet official website https://activity-net.org/ according to our provided video ID.

- Each video in the MECD dataset contains **4 to 11 events**, with a **minimum of 2 premise events exhibiting causal relations with the last one**.
- The annotations of training set: `train.json`, the annotations of testing set: `test.json`.
- The “duration”, “sentence”, and “timestamps” attributes of each video. (from ActivityNet Caption).
- A new attribute “**relation**” to **evaluate the video causal reasoning ability of the model** and support the supervised training.
- Full causal relation diagram annotations of the test set can be found at `test_complete.json`, and the additional attribute “**all_relation**” is introduced to **conduct complete causal graph reasoning** which is evaluated by the 'Average_Structural_Hamming_Distance (Ave SHD)' metric.



README.md

@@ -1,3 +1,5 @@
----
-license: mit
----
+---
+license: mit
+size_categories:
+- 1K<n<10K
+---