Update README.md

README.md

@@ -1,5 +1,9 @@
 ---
 language: en
+tags:
+- transformers
+- text-classification
+- taxonomy
 license: other
 license_name: link-attribution
 license_link: https://dejanmarketing.com/link-attribution/
@@ -16,6 +20,14 @@ This model is a hierarchical text classifier designed to categorize text into a
 - **Model Developers:** [DEJAN.AI](https://dejan.ai/)
 - **Model Type:** Hierarchical Text Classification
 - **Base Model:** [`albert/albert-base-v2`](https://huggingface.co/albert/albert-base-v2)
+- **Taxonomy Structure:** The model classifies text into a taxonomy with the following structure:
+    - **Level 1:** 21 unique classes
+    - **Level 2:** 193 unique classes
+    - **Level 3:** 1350 unique classes
+    - **Level 4:** 2205 unique classes
+    - **Level 5:** 1387 unique classes
+    - **Level 6:** 399 unique classes
+    - **Level 7:** 50 unique classes
 - **Model Architecture:**
     - **Level 1:** Standard sequence classification using `AlbertForSequenceClassification`.
     - **Levels 2-7:** Custom architecture (`TaxonomyClassifier`) where the ALBERT pooled output is concatenated with a one-hot encoded representation of the predicted ID from the previous level before being fed into a linear classification layer.
@@ -77,7 +89,7 @@ The model was trained on a dataset of 374,521 samples. Each row in the training
 Validation loss was used as the primary evaluation metric during training. The following validation loss trends were observed:
 
 - **Level 1, 2, and 3:** Showed a relatively rapid decrease in validation loss during training.
-- **Level 4:** Exhibited a slower decrease in validation loss, potentially due to the significant increase in the dimensionality of the parent ID one-hot encoding.
+- **Level 4:** Exhibited a slower decrease in validation loss, potentially due to the significant increase in the dimensionality of the parent ID one-hot encoding and the larger number of unique classes at this level.
 
 Further evaluation on downstream tasks is recommended to assess the model's practical performance.