Update README.md

README.md

@@ -1,3 +1,50 @@
----
-license: apache-2.0
----
+---
+language: en
+license: apache-2.0
+---
+
+# CodeRosetta
+## Pushing the Boundaries of Unsupervised Code Translation for Parallel Programming ([📃Paper](https://arxiv.org/abs/2410.20527), [🔗Website](https://coderosetta.com/)).
+
+
+CodeRosetta is an EncoderDecoder translation model. It supports the translation of C++, CUDA, and Fortran. \
+This version of the model is the base version of **C++-CUDA translation** without being fine-tuned.
+
+### How to use
+
+```python
+from transformers import AutoTokenizer, EncoderDecoderModel
+
+# Load the CodeRosetta model and tokenizer
+model = EncoderDecoderModel.from_pretrained('CodeRosetta/CodeRosetta_cpp2cuda_ft')
+tokenizer = AutoTokenizer.from_pretrained('CodeRosetta/CodeRosetta_cpp2cuda_ft')
+
+# Encode the input C++ Code
+input_cpp_code = "void add_100 ( int numElements , int * data ) { for ( int idx = 0 ; idx < numElements ; idx ++ ) { data [ idx ] += 100 ; } }"
+input_ids = tokenizer.encode(input_cpp_code, return_tensors="pt")
+
+# Set the start token to <CUDA>
+start_token = "<CUDA>" # If input is CUDA code, change the start token to <CPP>
+decoder_start_token_id = tokenizer.convert_tokens_to_ids(start_token)
+
+# Generate the CUDA code
+output = model.generate(
+    input_ids=input_ids, 
+    decoder_start_token_id=decoder_start_token_id,
+    max_length=256
+)
+
+# Decode and print the generated output
+generated_code= tokenizer.decode(output[0], skip_special_tokens=True)
+print(generated_code)
+```
+
+### BibTeX 
+
+```bibtex
+@inproceedings{coderosetta:neurips:2024,
+  title = {CodeRosetta: Pushing the Boundaries of Unsupervised Code Translation for Parallel Programming},
+  author = {TehraniJamsaz, Ali and Bhattacharjee, Arijit and Chen, Le and Ahmed, Nesreen K and Yazdanbakhsh, Amir and Jannesari, Ali},
+  booktitle = {NeurIPS},
+  year = {2024},
+}