Research map / Learning theory and algorithms / Prediction and decision losses
depyf: Open the Opaque Box of PyTorch Compiler for Machine Learning Researchers
Kaichao You, Runsheng Bai, Meng Cao, Jianmin Wang, Ion Stoica, Mingsheng Long
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use depyf instead of pycdc to decompile PyTorch compiler bytecode | bytecode decompilation · PyTorch compiler-generated bytecode tests | correctly decompiled test cases: 19.3%(27/140) -> 100%(140/140) | Improved |
| use depyf instead of pycdc on Python 3.11 bytecode | bytecode decompilation · Python 3.11 decompilation tests | correctly decompiled test cases: 67.1%(57/85) -> 100%(85/85) | Improved |
| use depyf instead of uncompyle6 on Python 3.8 bytecode | bytecode decompilation · Python 3.8 decompilation tests | correctly decompiled test cases: 91.8%(78/85) -> 100%(85/85) | Improved |
| use depyf instead of decompyle3 on Python 3.8 bytecode | bytecode decompilation · Python 3.8 decompilation tests | correctly decompiled test cases: 90.6%(77/85) -> 100%(85/85) | Improved |
Closest papers by meaning
- Scaling Tractable Probabilistic Circuits: A Systems Perspective (ICML 2024 · similarity 0.79)
- $\bf{\Phi}_\textrm{Flow}$: Differentiable Simulations for PyTorch, TensorFlow and Jax (ICML 2024 · similarity 0.79)
- LILO: Learning Interpretable Libraries by Compressing and Documenting Code (ICLR 2024 · similarity 0.78)
- Pfeife: Automatic Pipeline Parallelism for PyTorch (ICML 2025 · similarity 0.78)
- CodeRosetta: Pushing the Boundaries of Unsupervised Code Translation for Parallel Programming (NeurIPS 2024 · similarity 0.78)
- KernelBench: Can LLMs Write Efficient GPU Kernels? (ICML 2025 · similarity 0.78)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
