Skip to content

Actions: fattorib/ZeRO-transformer

Actions

Tests

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
153 workflow runs
153 workflow runs

Filter by Event

Loading

Filter by Status

Loading

Filter by Branch

Loading

Filter by Actor

Loading
remove one layer
Tests #1098: Commit f147adb pushed by fattorib
June 4, 2023 18:38 3m 38s pjit-tensor-parallel
June 4, 2023 18:38 3m 38s
more layer test
Tests #1097: Commit 84ba220 pushed by fattorib
June 4, 2023 18:22 4m 23s pjit-tensor-parallel
June 4, 2023 18:22 4m 23s
residual connection fixes + lower peak LR
Tests #1096: Commit 1614e32 pushed by fattorib
June 4, 2023 18:21 3m 40s pjit-tensor-parallel
June 4, 2023 18:21 3m 40s
halve total eval steps
Tests #1095: Commit efc8388 pushed by fattorib
June 4, 2023 14:45 3m 28s pjit-tensor-parallel
June 4, 2023 14:45 3m 28s
revert config
Tests #1094: Commit a97775e pushed by fattorib
June 4, 2023 14:45 3m 29s pjit-tensor-parallel
June 4, 2023 14:45 3m 29s
slightly smaller 2.6B
Tests #1093: Commit 4c378bd pushed by fattorib
June 4, 2023 13:58 3m 30s pjit-tensor-parallel
June 4, 2023 13:58 3m 30s
config setup for 2.7B
Tests #1092: Commit b097a51 pushed by fattorib
June 4, 2023 13:29 3m 34s pjit-tensor-parallel
June 4, 2023 13:29 3m 34s
enable bf16 casting
Tests #1090: Commit 668d029 pushed by fattorib
June 4, 2023 13:08 3m 39s pjit-tensor-parallel
June 4, 2023 13:08 3m 39s
disable act print in GPT
Tests #1088: Commit 7c0c6f2 pushed by fattorib
June 4, 2023 12:49 3m 59s pjit-tensor-parallel
June 4, 2023 12:49 3m 59s
print a single data batch
Tests #1087: Commit b5e3c87 pushed by fattorib
June 4, 2023 12:47 4m 13s pjit-tensor-parallel
June 4, 2023 12:47 4m 13s
check layer right after embed
Tests #1086: Commit 17d5aa3 pushed by fattorib
June 4, 2023 12:44 3m 36s pjit-tensor-parallel
June 4, 2023 12:44 3m 36s
debug out bf16 casting
Tests #1085: Commit 0b0b715 pushed by fattorib
June 4, 2023 12:41 4m 18s pjit-tensor-parallel
June 4, 2023 12:41 4m 18s
remove ALiBi head shard
Tests #1084: Commit 27c85f5 pushed by fattorib
June 4, 2023 12:36 3m 50s pjit-tensor-parallel
June 4, 2023 12:36 3m 50s
move further into model for debug
Tests #1083: Commit e599a6a pushed by fattorib
June 4, 2023 12:32 3m 29s pjit-tensor-parallel
June 4, 2023 12:32 3m 29s
loss
Tests #1082: Commit dd68768 pushed by fattorib
June 4, 2023 12:29 3m 40s pjit-tensor-parallel
June 4, 2023 12:29 3m 40s
debug in loss
Tests #1081: Commit 439f56f pushed by fattorib
June 4, 2023 12:26 4m 7s pjit-tensor-parallel
June 4, 2023 12:26 4m 7s
debug loss
Tests #1080: Commit 89de101 pushed by fattorib
June 4, 2023 12:23 3m 39s pjit-tensor-parallel
June 4, 2023 12:23 3m 39s
missing psum
Tests #1079: Commit c80ee4f pushed by fattorib
June 4, 2023 12:17 3m 43s pjit-tensor-parallel
June 4, 2023 12:17 3m 43s
convert logits to f32 earlier
Tests #1078: Commit a77988f pushed by fattorib
June 4, 2023 12:03 3m 43s pjit-tensor-parallel
June 4, 2023 12:03 3m 43s
smaller model for debug
Tests #1077: Commit a24b7ed pushed by fattorib
June 4, 2023 12:01 4m 18s pjit-tensor-parallel
June 4, 2023 12:01 4m 18s
print
Tests #1076: Commit 79728cd pushed by fattorib
June 4, 2023 11:58 4m 8s pjit-tensor-parallel
June 4, 2023 11:58 4m 8s
arg reordering
Tests #1075: Commit 129eb3f pushed by fattorib
June 4, 2023 11:47 4m 18s pjit-tensor-parallel
June 4, 2023 11:47 4m 18s
fix head division for sharding
Tests #1074: Commit d3406bf pushed by fattorib
June 4, 2023 11:44 3m 37s pjit-tensor-parallel
June 4, 2023 11:44 3m 37s