Aleksandra Еliseeva

Models' contexts are getting so big that they can (and often should!) include an entire repository, while we are still evaluating them on methods and files. That's why we created Long Code Arena. Pre-print: arxiv.org/abs/2406.11612 Datasets: huggingface.co/spaces/JetBrai… Details in🧵!

Our dataset, models and all the code are available online! 💛 💻Repository: github.com/JetBrains-Rese… 🤗Dataset: huggingface.co/datasets/JetBr… 🤗Models: #available-checkpoints" target="_blank" rel="nofollow noopener">huggingface.co/JetBrains-Rese… 🌐We also have an accompanying website: commit-chronicle.github.io 9/9

We also compare results with and without frequent data filtering steps. Our experiments show that the results on filtered subset are higher than on random subset of the same size, while results for commits that fit neither of the filters are significantly lower. 8/9

We published a preprint for our paper "From Commit Message Generation to History-Aware Commit Message Completion" from the main track of @ASE_conf! 📜 arxiv.org/abs/2308.07655 📜 Thanks to my amazing co-authors: Yaroslav, @egor_bb, @areyde, @danny_dig and @timofeybryksin 🧵 1/9

I am honored to share that my first paper "From Commit Message Generation to History-Aware Commit Message Completion" has been accepted to the main track of @ASE_conf! Congratulations to Yaroslav, @egor_bb, @areyde, @danny_dig, and @timofeybryksin💛 Will share the pre-print soon!