Skip to content

Actions: NVIDIA/TransformerEngine

Deploy nightly docs

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
667 workflow runs
667 workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

[PyTorch] Improve get_qkv_layout (#1214)
Deploy nightly docs #662: Commit 5b6546c pushed by cyanguwa
October 9, 2024 16:48 1m 12s main
October 9, 2024 16:48 1m 12s
[PyTorch] Add documentation for FP8 attention checkpointing (#1223)
Deploy nightly docs #661: Commit 2d87552 pushed by cyanguwa
October 9, 2024 16:47 1m 15s main
October 9, 2024 16:47 1m 15s
[PyTorch] Debug dtype casting in operation-based API (#1202)
Deploy nightly docs #660: Commit 5b89f1a pushed by timmoon10
October 9, 2024 03:58 1m 14s main
October 9, 2024 03:58 1m 14s
[PyTorch] Miscellaneous fixes for FA3 attention (#1174)
Deploy nightly docs #659: Commit e762592 pushed by cyanguwa
October 8, 2024 18:06 1m 3s main
October 8, 2024 18:06 1m 3s
Fix cuDNN sliding window size (#1212)
Deploy nightly docs #658: Commit c3b3cd2 pushed by cyanguwa
October 7, 2024 21:31 1m 43s main
October 7, 2024 21:31 1m 43s
Hierarchical CP implementation (Ulysses + Ring) (#1209)
Deploy nightly docs #657: Commit c24a4c4 pushed by cyanguwa
October 7, 2024 21:14 1m 3s main
October 7, 2024 21:14 1m 3s
Tests for distributed (#1196)
Deploy nightly docs #656: Commit 60f738f pushed by ptrendx
October 7, 2024 16:44 1m 12s main
October 7, 2024 16:44 1m 12s
[PyTorch] remove duplicate code (#1215)
Deploy nightly docs #655: Commit f8eb799 pushed by ksivaman
October 6, 2024 15:18 1m 16s main
October 6, 2024 15:18 1m 16s
[PyTorch] Minor optimizations to reduce CPU overheads in modules (#1191)
Deploy nightly docs #654: Commit 9d976bc pushed by timmoon10
October 4, 2024 03:13 1m 7s main
October 4, 2024 03:13 1m 7s
[PyTorch] Move block_table argument to FA varlen function (#1222)
Deploy nightly docs #653: Commit 10cceae pushed by cyanguwa
October 3, 2024 15:58 1m 13s main
October 3, 2024 15:58 1m 13s
Removed the unused options from GroupedLinear docs and fixed the bug …
Deploy nightly docs #652: Commit fb74961 pushed by ksivaman
October 1, 2024 02:33 1m 9s main
October 1, 2024 02:33 1m 9s
[PyTorch] Fix distributed testing (#1219)
Deploy nightly docs #651: Commit 46075b9 pushed by ksivaman
October 1, 2024 02:33 1m 6s main
October 1, 2024 02:33 1m 6s
[PyTorch] Add pool argument to make_graphed_callable (#1218)
Deploy nightly docs #650: Commit 728c558 pushed by ksivaman
October 1, 2024 01:00 1m 0s main
October 1, 2024 01:00 1m 0s
Fix CP unit test on A100 and L40s (#1211)
Deploy nightly docs #649: Commit 7b152a8 pushed by xrennvidia
September 27, 2024 18:56 1m 3s main
September 27, 2024 18:56 1m 3s
[PyTorch] Fix detection of 3 in 3hd/h3d layouts (#1187)
Deploy nightly docs #648: Commit 8a1b7ee pushed by ptrendx
September 27, 2024 18:33 1m 31s main
September 27, 2024 18:33 1m 31s
[PyTorch] Add GroupedLinear to the docs and fix typos (#1206)
Deploy nightly docs #647: Commit c4a5cb8 pushed by ksivaman
September 27, 2024 17:48 1m 0s main
September 27, 2024 17:48 1m 0s
fix NVTE_UB_WITH_MPI read (#1194)
Deploy nightly docs #646: Commit 209b8e5 pushed by erhoo82
September 25, 2024 04:24 1m 10s main
September 25, 2024 04:24 1m 10s
Update list of CI users (#1203)
Deploy nightly docs #645: Commit a44cb72 pushed by ksivaman
September 24, 2024 18:59 1m 13s main
September 24, 2024 18:59 1m 13s
Allow to pass architectures like 90a, without being overriden (#1178)
Deploy nightly docs #644: Commit 99af5c0 pushed by timmoon10
September 24, 2024 18:01 1m 7s main
September 24, 2024 18:01 1m 7s
Update list of CI users (#1198)
Deploy nightly docs #643: Commit a68acd7 pushed by timmoon10
September 23, 2024 18:12 1m 18s main
September 23, 2024 18:12 1m 18s
Restore compatibility with Python 3.8 (#1189)
Deploy nightly docs #642: Commit 0c74535 pushed by ptrendx
September 20, 2024 23:05 1m 4s main
September 20, 2024 23:05 1m 4s
Allow downloading of model weights automatically (#1172)
Deploy nightly docs #641: Commit 195d703 pushed by ptrendx
September 20, 2024 20:38 1m 13s main
September 20, 2024 20:38 1m 13s
[PyTorch] Relax the contiguous check for flash attention (#1176)
Deploy nightly docs #640: Commit 0ee5ccd pushed by cyanguwa
September 19, 2024 00:55 1m 22s main
September 19, 2024 00:55 1m 22s
Expose rotary_base as an arg instead of hardcoding (#944)
Deploy nightly docs #639: Commit c0caadb pushed by ptrendx
September 18, 2024 22:31 2m 7s main
September 18, 2024 22:31 2m 7s
[PyTorch] Check network interface name when initializing Userbuffers …
Deploy nightly docs #638: Commit 841634c pushed by denera
September 18, 2024 18:09 1m 8s main
September 18, 2024 18:09 1m 8s
ProTip! You can narrow down the results and go further in time using created:<2024-09-18 or the other filters available.