init README.md

README.md

@@ -0,0 +1,47 @@
+---
+language: en
+tags:
+- ABSA
+- aspect-based-sentiment-analysis
+- pytorch
+datasets:
+- semeval2014
+widget:
+- text: "[CLS] The appearance is very nice, but the battery life is poor. [SEP] appearance [SEP] "
+---
+
+# Note
+
+BERT based ABSA baseline, based on https://github.com/avinashsai/BERT-Aspect *BERT LSTM* implementation.
+
+Code for the paper "Utilizing BERT Intermediate Layers for Aspect Based Sentiment Analysis and Natural Language Inference" https://arxiv.org/pdf/2002.04815.pdf.
+
+When using the widget, please splice the aspect vocabulary behind the original sentence by [SEP].
+
+# Usage
+
+```python
+from transformers import AutoTokenizer, AutoModelForSequenceClassification, TextClassificationPipeline
+
+MODEL = "tezign/BERT-LSTM-based-ABSA"
+
+tokenizer = AutoTokenizer.from_pretrained(MODEL)
+
+model = AutoModelForSequenceClassification.from_pretrained(MODEL, trust_remote_code=True)
+
+classifier = TextClassificationPipeline(model=model, tokenizer=tokenizer)
+
+result = classifier([
+    {"text": "The appearance is very nice, but the battery life is poor", "text_pair": "appearance"},
+    {"text": "The appearance is very nice, but the battery life is poor", "text_pair": "battery"}
+],
+    function_to_apply="softmax")
+
+print(result)
+
+"""
+print result
+>> [{'label': 'positive', 'score': 0.9129462838172913}, {'label': 'negative', 'score': 0.8834680914878845}]
+"""
+
+```