Ongoing research training transformer models at scale
This is an exact mirror of the Megatron project, hosted at https://github.com/NVIDIA/Megatron-LM.
SourceForge is not affiliated with Megatron.
Megatron is a trademark of NVIDIA CORPORATION.
Changelog Details
- beep boop 🤖: Bumping Megatron Core to v0.18.1 by @nvidia-megatron-lm-release-bot[bot] :: PR: [#5439]
- build: add fast-hadamard-transform and point flash_mla at nv_dev by @ko3n1g :: PR: [#5447]
- cp: `build: install flash_mla from source in the CI image (5481)` into `core_r0.18.0` by @ko3n1g :: PR: [#5484]
- cp: `Disable TE cross entropy loss fusion (5115)` into `core_r0.18.0` by @ko3n1g :: PR: [#5144]
- chore: bump megatron-core to v0.18.2 by @ko3n1g :: PR: [#5499]
- cp: `Move TE cross entropy guard to training args (5162)` into `core_r0.18.0` by @ko3n1g :: PR: [#5185]
- cp: `build: bump transformer-engine to release_v2.16.post` into `core_r0.18.0` by @ko3n1g :: PR: [#5518]
- beep boop 🤖: Bumping Megatron Core to v0.18.1 [skip ci] by @github-actions[bot]
- build: add fast-hadamard-transform and point flash_mla at nv_dev (#5447) by @ko3n1g
- cp: `build: install flash_mla from source in the CI image (5481)` into `core_r0.18.0` (#5484) by @ko3n1g
- cp: `Disable TE cross entropy loss fusion (5115)` into `core_r0.18.0` (#5144) by @ko3n1g
- chore: bump megatron-core to v0.18.2 (#5499) by @ko3n1g
- cp: `Move TE cross entropy guard to training args (5162)` into `core_r0.18.0` (#5185) by @ko3n1g
- cp: `build: bump transformer-engine to release_v2.16.post` into `core_r0.18.0` (#5518) by @ko3n1g
- chore(beep boop 🤖): Bump (core_r0.18.0) (2026-06-29) by @github-actions[bot]
- chore(beep boop 🤖): Bump (core_r0.18.0) (2026-07-20) by @github-actions[bot]
Source: README.md, updated 2026-07-20
Other Useful Business Software
MongoDB Atlas runs apps anywhere
Deploy in 115+ regions with the modern database for every enterprise.
MongoDB Atlas gives you the freedom to build and run modern applications anywhere—across AWS, Azure, and Google Cloud. With global availability in over 115 regions, Atlas lets you deploy close to your users, meet compliance needs, and scale with confidence across any geography.
Start Free
Save Up to 91% on Cloud Compute With Spot VMs
Automatic sustained-use discounts. One free VM per month. No negotiation needed.
Run batch jobs at 60-91% off with Spot VMs. Long-running workloads get automatic discounts with sustained use.