Doc fix: TORCH_ROCM_AOTRITON_ENABLE_EXPERIMENTAL not necessary to use Triton kernel patches (#2343 )

* removing note about TORCH_ROCM_AOTRITON_ENABLE_EXPERIMENTAL * suggest using TORCH_ROCM_AOTRITON_ENABLE_EXPERIMENTAL for memory efficient attn
bump dev version (#2342 )
2025-02-18 10:06:31 -05:00 · 2025-02-18 04:30:59 -05:00
2 changed files with 2 additions and 2 deletions
--- a/docs/lora_optims.qmd
+++ b/docs/lora_optims.qmd
@@ -82,7 +82,7 @@ lora_o_kernel: true
 ## Requirements

 - One or more NVIDIA or AMD GPUs (in order to use the Triton kernels)
-    - AMD can be used with experimental Triton support by setting the environment variable `TORCH_ROCM_AOTRITON_ENABLE_EXPERIMENTAL=1`
+    - Note: Set `TORCH_ROCM_AOTRITON_ENABLE_EXPERIMENTAL=1` to enable [memory-efficient attention on AMD GPUs](https://github.com/ROCm/aotriton/issues/16#issuecomment-2346675491)
 - Targeted LoRA adapters cannot use Dropout
    - This may limit model expressivity / cause overfitting
 - Targeted LoRA adapters cannot have bias terms
--- a/src/axolotl/init.py
+++ b/src/axolotl/init.py
@@ -4,4 +4,4 @@ import pkgutil

 __path__ = pkgutil.extend_path(__path__, __name__)  # Make this a namespace package

-__version__ = "0.7.0"
+__version__ = "0.8.0.dev0"
Author	SHA1	Message	Date
Dan Saunders	c3d4f6e295	Doc fix: TORCH_ROCM_AOTRITON_ENABLE_EXPERIMENTAL not necessary to use Triton kernel patches (#2343 ) * removing note about TORCH_ROCM_AOTRITON_ENABLE_EXPERIMENTAL * suggest using TORCH_ROCM_AOTRITON_ENABLE_EXPERIMENTAL for memory efficient attn	2025-02-18 10:06:31 -05:00
Wing Lian	7fa690fac8	bump dev version (#2342 )	2025-02-18 04:30:59 -05:00