create first README

README.md

@@ -1,14 +1,9 @@
-This model was trained with MLSUM (Turkish language) summarization datasets where I fine-tuned google/mt5-small using [SimpleT5](https://github.com/Shivanandroy/simpleT5) library.
-The first results are not promising may be due to using small check-points. I will work on it for improvements!
-
-The code piece for training
+This checkpoint is trained with the Turkish part of the MLSUM dataset where google/mt5 PLM is fine-tuned. [SimpleT5](https://github.com/Shivanandroy/simpleT5) library is used to fine-tune. Here is the code snippet for training
 
 ```
-from simplet5 import SimpleT5
 model = SimpleT5()
 model.from_pretrained("mt5","google/mt5-small")
 
-# train
 model.train(train_df=train2, # pandas dataframe with 2 columns: source_text & target_text
             eval_df=validation2, # pandas dataframe with 2 columns: source_text & target_text
             source_max_token_len = 512, 
@@ -21,4 +16,3 @@ model.train(train_df=train2, # pandas dataframe with 2 columns: source_text & ta
             precision = 32
 )
 ```
-