CuPy

Drop-in replacement of NumPy and SciPy for GPU - Follow us for the latest updates!
🫶 Support us via GitHub Sponsors:

[CuPy @CuPy_Team]

Announcing CuPy v14! 🚀

🔹 NumPy v2 semantics
🔹 CUDA Pip Wheels support
🔹 bfloat16 & structured dtype
🔹 Enhanced NumPy/SciPy API coverage

Read more on our blog for the full details! 👇

[CuPy @CuPy_Team]

@NumFOCUS 🚀

[CuPy @CuPy_Team]

@PyTorch @nvidia 🙌

[PyTorch @PyTorch]

Major ecosystem libraries like PyTorch and CuPy now build on the shared cuda.core foundation, reducing version conflicts and interop bugs for developers.

With CUDA 13.3, @nvidia released CUDA Python 1.0, the libraries and tools that give you the full CUDA platform from Python. Using standard CUDA context, devices, streams, and memory are shared with the rest of the Python GPU ecosystem, so your own kernels can run against CuPy arrays or PyTorch tensors without copying data.

Read the full blog post: https://t.co/UNoXigEQBd

[CuPy @CuPy_Team]

https://t.co/uBfitBvd11

#CuPy #Python #DataScience #CUDA #NumPy #SciPy

[CuPy @CuPy_Team]

CuPy v14.2 is out! 🎉
Highlights:
🔹 SciPy-compatible sparse array classes
🔹 Windows on Arm (RTX Spark) support
🔹 Initial integration with nvmath-python

Check the release notes for details 👇

[CuPy @CuPy_Team]

https://t.co/OLqT5QTZye

[CuPy @CuPy_Team]

CuPy v14.1.0 is out! 🚀 Large sparse matrices, Python 3.14t (free-threaded) support, structured dtypes, CUDA 13.2/NCCL 2.29, new APIs, and more. Check out the release notes for details! 🧵

[mohit @mohitwt_]

https://t.co/uLqz88Ve0n

[CuPy @CuPy_Team]

https://t.co/Rbk3ltaqJ0

[CuPy @CuPy_Team]

Announcing CuPy v14! 🚀

🔹 NumPy v2 semantics
🔹 CUDA Pip Wheels support
🔹 bfloat16 & structured dtype
🔹 Enhanced NumPy/SciPy API coverage

Read more on our blog for the full details! 👇

[CuPy @CuPy_Team]

https://t.co/uUw79UQJpk

[CuPy @CuPy_Team]

🚀 CuPy v14.0.0rc1 is here!

Highlights include CUDA PyPI package compatibility (no system-wide CUDA Toolkit installation required) and AMD ROCm 7 support.

Test your workload with our pre-compiled packages and report any feedback! See the release notes for the full details 👇

[CuPy @CuPy_Team]

Release notes: https://t.co/gq8gJMV20u

[CuPy @CuPy_Team]

CuPy v13.6 is out, now with CUDA 13 support! 🚀
Install with: pip install cupy-cuda13x

[CuPy @CuPy_Team]

Recording is now available on YouTube! https://t.co/7fRTgCFUGk

[CuPy @CuPy_Team]

⚡️A talk on CuPy will be presented at the SciPy 2025 conference! 🎉
Leo Fang, a long-time contributor, will share his personal story on #CuPy's evolution and future.
https://t.co/R1lcHUNSQu
#SciPy2025

[CuPy @CuPy_Team]

Also, there are more sessions and tutorials highlighting CuPy at SciPy 2025. Explore the full schedule at https://t.co/ypbYJYB4dn!

[CuPy @CuPy_Team]

https://t.co/GdVVbizGSP

[CuPy @CuPy_Team]

📢CuPy v13.5 is out! CuPy now supports CUDA 12.9, ROCm 6.4, and NumPy 2.3. This release also brings Unified Memory Programming support for NVIDIA’s Grace Hopper systems.

Check out the full release notes 🧵

[CuPy @CuPy_Team]

Thanks for featuring, @StarHistoryHQ! https://t.co/Pd7sGGuLDy

[Star History @StarHistoryHQ]

10.1K ⭐️ CuPy is a NumPy/SciPy-compatible array library 📚 that enables GPU-accelerated computing in Python by serving as a drop-in replacement for existing NumPy/SciPy code. @CuPy_Team

https://t.co/d0IYUmjW2h

#starhistory #GitHub #OpenSource

[CuPy @CuPy_Team]

One more thing - CuPy has just reached 10,000 GitHub stars! ⭐ A huge thank you to all our users, contributors, and everyone who has supported us on this incredible journey!

If you haven't yet, now's a great time to add your star 😉
https://t.co/yYS2EUYm7K