Create README.md

README.md

@@ -0,0 +1,28 @@
+An Arabic abstractive text summarization model.
+A fine-tuned AraT5 model on a dataset that consists of 86,523 paragraph-summary pairs.
+
+More details on the fine-tuning of this model will be released later.
+
+The model can be used as follows:
+```python
+from arabert.preprocess import ArabertPreprocessor
+
+model_name="malmarjeh/t5-arabic-text-summarization"
+arabert_prep = ArabertPreprocessor(model_name=model_name)
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+model = AutoModelForSeq2SeqLM.from_pretrained(model_name)
+pipeline = pipeline("text2text-generation",model=model,tokenizer=tokenizer)
+
+text = "ولن نبالغ إذا قلنا إن هاتف أو كمبيوتر المكتب في زمننا هذا ضروري"
+preprocessor = ArabertPreprocessor(model_name="")
+preprocessor.preprocess(text)
+result = pipeline(text,
+            pad_token_id=tokenizer.eos_token_id,
+            num_beams=3,
+            repetition_penalty=3.0,
+            max_length=200,
+            length_penalty=1.0,
+            no_repeat_ngram_size = 3)[0]['generated_text']
+result
+>>>"و+ لن نبالغ إذا قل +نا إن هاتف أو كمبيوتر ال+ مكتب في زمن +نا هذا ضروري"
+```