tiled_mlp supports single gpu (#2891)
Some checks failed
ci-cd / build-axolotl (<nil>, 126, 12.6.3, 3.11, 2.6.0) (push) Has been cancelled
ci-cd / build-axolotl (<nil>, 126, 12.6.3, 3.11, 2.7.1) (push) Has been cancelled
ci-cd / build-axolotl (<nil>, 128, 12.8.1, 3.11, 2.7.1) (push) Has been cancelled
ci-cd / build-axolotl (vllm, 126, 12.6.3, 3.11, 2.7.0) (push) Has been cancelled
publish pypi / Create Release (push) Has been cancelled
ci-cd / build-axolotl-cloud (<nil>, 126, 12.6.3, 3.11, 2.7.0) (push) Has been cancelled
ci-cd / build-axolotl-cloud (<nil>, 126, 12.6.3, 3.11, 2.7.1) (push) Has been cancelled
ci-cd / build-axolotl-cloud (<nil>, 126, 12.6.3, true, 3.11, 2.6.0) (push) Has been cancelled
ci-cd / build-axolotl-cloud (<nil>, 128, 12.8.1, 3.11, 2.7.1) (push) Has been cancelled
ci-cd / build-axolotl-cloud-no-tmux (<nil>, 126, 12.6.3, 3.11, 2.6.0) (push) Has been cancelled
publish pypi / Upload release to PyPI (push) Has been cancelled

* tiled_mlp supports single gpu

* use checkpoint offloading for arctic training

* patch torch checkpoint too

* support for single gpu zero3

* add linkback to where it was copied from
This commit is contained in:
Wing Lian
2025-07-09 12:48:22 -04:00
parent c6d69d5c1b
commit c620a218b8
7 changed files with 218 additions and 7 deletions

View File

@@ -66,6 +66,15 @@ Start from Stage 1 -> Stage 2 -> Stage 3.
:::
::: {.callout-tip}
Using ZeRO Stage 3 with Single-GPU training
ZeRO Stage 3 can be used for training on a single GPU by manually setting the environment variables:
`WORLD_SIZE=1 LOCAL_RANK=0 MASTER_ADDR=0.0.0.0 MASTER_PORT=29500`
:::
## FSDP {#sec-fsdp}
### Basic FSDP Configuration {#sec-fsdp-config}