Update README.md

README.md

@@ -54,42 +54,35 @@ You can use this model directly with a pipeline for masked language modeling:
   'sequence': 'ከሀገራቸው ከኢትዮጵያ ከወጡ ግማሽ ምዕተ ዓመታትን ተቆጥሯል ።'}]
 ```
 
-# Fine-tuning
+# Finetuning
 
-The following github repository contains a [notebook](https://github.com/rasyosef/amharic-news-category-classification/blob/main/%5Bbert-mini-amharic%5D%20Amharic%20News%20Category%20Classification.ipynb) that fine-tunes this model for an Amharic text classification task. 
+This model was finetuned and evaluated on the following amharic nlp tasks
 
-https://github.com/rasyosef/amharic-news-category-classification
+- Sentiment Classification
+  - Dataset: [amharic-sentiment](https://huggingface.co/datasets/rasyosef/amharic-sentiment)
+- Named Entity Recognition
+  - Dataset: [amharic-named-entity-recognition](https://huggingface.co/datasets/rasyosef/amharic-named-entity-recognition)
+- News Category Classification
+  - Dataset: [amharic-news-category-classification](https://github.com/rasyosef/amharic-news-category-classification)
 
-#### Fine-tuned Model Performance
-Since this is a multi-class classification task, the reported precision, recall, and f1 metrics are macro averages.
+### Finetuned Model Performance
+The reported F1 scores are macro averages.
 
-|Model|Size(# params)|Accuracy|Precision|Recall|F1|
-|-----|--------------|--------|---------|------|--|
-|bert-mini-amharic|9.67M|0.87|0.83|0.83|0.83|
-|bert-small-amharic|25.7M|0.89|0.86|0.87|0.86|
-|xlm-roberta-base|279M|0.9|0.88|0.88|0.88|
-
-### Sentiment Classification
-The model was finetuned on the [amharic-sentiment](https://huggingface.co/datasets/rasyosef/amharic-sentiment) 
-dataset to classify the given text as having `positive` or `negative` sentiment
+|Model|Size (# params)| Perplexity|Sentiment (F1)| Named Entity Recognition (F1)|
+|-----|---------------|-----------|--------------|------------------------------|
+|bert-medium-amharic|40.5M|13.74|0.83|0.68|
+|bert-small-amharic|27.8M|15.96|0.83|0.68|
+|**bert-mini-amharic**|**10.7M**|**22.42**|**0.81**|**0.64**|
+|bert-tiny-amharic|4.18M|71.52|0.79|0.54|
+|xlm-roberta-base|279M||0.83|0.73|
+|am-roberta|443M||0.82|0.69|
 
-|Model|Size (# params)| Accuracy | Precision | Recall | F1 |
-| --- | ------------- | -------- | --------- | ------ | -- |
-|bert-medium-amharic|40.5M|0.83|0.83|0.82|0.83|
-|bert-small-amharic|27.8M|0.83|0.83|0.82|0.83|
-|**bert-mini-amharic**|10.7M|0.81|0.81|0.81|0.81|
-|bert-tiny-amharic|4.18M|0.79|0.79|0.79|0.79|
-|xlm-roberta-base|279M|0.83|0.83|0.83|0.83|
-|am-roberta|443M|0.82|0.83|0.82|0.82|
+### Amharic News Category Classification
 
-### Named Entity Recognition
-The model was finetuned on the [amharic-named-entity-recognition](https://huggingface.co/datasets/rasyosef/amharic-named-entity-recognition) dataset.
+The following github repository contains a [notebook](https://github.com/rasyosef/amharic-news-category-classification/blob/main/%5Bbert-mini-amharic%5D%20Amharic%20News%20Category%20Classification.ipynb) that fine-tunes this model for an Amharic text classification task using the [amharic-news-category-classification](https://github.com/rasyosef/amharic-news-category-classification) dataset.
 
-|Model|Size (# params)| Precision | Recall | F1 |
-| --- | ------------- | --------- |------- | -- |
-|bert-medium-amharic|40.5M|0.64|0.73|0.68|
-|bert-small-amharic|27.8M|0.64|0.72|0.68|
-|**bert-mini-amharic**|10.7M|0.60|0.67|0.64|
-|bert-tiny-amharic|4.18M|0.50|0.59|0.54|
-|xlm-roberta-base|279M|0.69|0.79|0.73|
-|am-roberta|443M|0.67|0.72|0.69|
+|Model|Size(# params)|Accuracy|Precision|Recall|F1|
+|-----|--------------|--------|---------|------|--|
+|bert-small-amharic|25.7M|0.89|0.86|0.87|0.86|
+|**bert-mini-amharic**|9.67M|0.87|0.83|0.83|0.83|
+|xlm-roberta-base|279M|0.9|0.88|0.88|0.88|