v1.1.0.post2

Full Changelog: v1.1.0.post1...v1.1.0.post2

v1.1.0.post1

Full Changelog: v1.1.0...v1.1.0.post1

v1.1.0

What's Changed

• Bug fix/fix had 12 by @hychiang-git in #5
• Feats/add had 40 by @hychiang-git in #6
• Don't compile for sm < 75 if nvcc >= 13.0 by @kevalmorabia97 in #17
• fix: mising import by @xxrjun in #18
• fix: correct dimension checks and log_N condition by @xxrjun in #19
• Update cpp extension with blackwell support by @lessw2020 in #15
• 12.6.1 doesn't support sm100, so move version to 12.8.1 by @gongwei-130 in #20
• Update README with installation instructions by @Huixxi in #22
• fix sm100 cuda bare metal version condition by @dianyo in #23
• Support to set nvcc threads by environment variable NVCC_THREADS by @acelyc111 in #24

New Contributors

• @hychiang-git made their first contribution in #5
• @kevalmorabia97 made their first contribution in #17
• @xxrjun made their first contribution in #18
• @lessw2020 made their first contribution in #15
• @gongwei-130 made their first contribution in #20
• @Huixxi made their first contribution in #22
• @dianyo made their first contribution in #23
• @acelyc111 made their first contribution in #24

Full Changelog: v1.0.4...v1.1.0

v1.0.4.post1

[CI] Exclude Pytorch 2.1.2 python 3.7

v1.0.4

Bump to v1.0.4

v1.0.3.post3

[CI] Install typing-extensions before installing pytorch pre

v1.0.3.post2

[CI] Install Pytorch with --no-deps

v1.0.3.post1

Use URL to install torch.20231106dev

v1.0.3

Bump to v1.0.3

v1.0.1

Implement Hadamard transform for dim = 20 * 2^k and 28 * 2^k