From d1fd50581302fcde27f1f78f749aab22cfe45dc4 Mon Sep 17 00:00:00 2001
From: Dan Saunders <danjsaund@gmail.com>
Date: Sat, 30 Aug 2025 00:42:38 -0400
Subject: [PATCH] update

---
 README.md             | 12 ++++--------
 cicd/Dockerfile.jinja |  4 ++--
 docker/Dockerfile     |  4 ++--
 docs/installation.qmd |  7 ++++---
 pyproject.toml        |  8 ++------
 5 files changed, 14 insertions(+), 21 deletions(-)

diff --git a/README.md b/README.md
index 8ff5e361f..b87103a35 100644
--- a/README.md
+++ b/README.md
@@ -72,8 +72,7 @@ Features:
 **Requirements**:
 
 - NVIDIA GPU (Ampere or newer for `bf16` and Flash Attention) or AMD GPU
-- Python 3.11
-- PyTorch ≥2.6.0
+- Python 3.11+
 
 ### Google Colab
 
@@ -87,11 +86,8 @@ Features:
 # Install uv
 curl -LsSf https://astral.sh/uv/install.sh | sh
 
-# Create virtualenv
-uv venv
-
-# Install axolotl
-uv pip install --no-build-isolation axolotl[flash-attn,deepspeed]
+# Add axolotl to your project
+uv add axolotl
 
 # Download example axolotl configs, deepspeed configs
 axolotl fetch examples
@@ -102,7 +98,7 @@ axolotl fetch deepspeed_configs  # OPTIONAL
 
 ```bash
 pip3 install -U packaging==23.2 setuptools==75.8.0 wheel ninja
-pip3 install --no-build-isolation axolotl[flash-attn,deepspeed]
+pip3 install --no-build-isolation axolotl
 
 # Download example axolotl configs, deepspeed configs
 axolotl fetch examples
diff --git a/cicd/Dockerfile.jinja b/cicd/Dockerfile.jinja
index d294460d7..f05be6857 100644
--- a/cicd/Dockerfile.jinja
+++ b/cicd/Dockerfile.jinja
@@ -33,9 +33,9 @@ RUN if [ "$NIGHTLY_BUILD" = "true" ] ; then \
 
 RUN uv pip install packaging==23.2 setuptools==75.8.0
 RUN if [ "$AXOLOTL_EXTRAS" != "" ] ; then \
-        uv pip install --no-build-isolation -e .[deepspeed,flash-attn,ring-flash-attn,optimizers,ray,$AXOLOTL_EXTRAS] $AXOLOTL_ARGS; \
+        uv pip install --no-build-isolation -e .[ring-flash-attn,optimizers,ray,$AXOLOTL_EXTRAS] $AXOLOTL_ARGS; \
     else \
-        uv pip install --no-build-isolation -e .[deepspeed,flash-attn,ring-flash-attn,optimizers,ray] $AXOLOTL_ARGS; \
+        uv pip install --no-build-isolation -e .[ring-flash-attn,optimizers,ray] $AXOLOTL_ARGS; \
     fi
 
 RUN python scripts/unsloth_install.py | sh
diff --git a/docker/Dockerfile b/docker/Dockerfile
index 2c64364ac..dcf2637c4 100644
--- a/docker/Dockerfile
+++ b/docker/Dockerfile
@@ -22,9 +22,9 @@ WORKDIR /workspace/axolotl
 
 # If AXOLOTL_EXTRAS is set, append it in brackets
 RUN if [ "$AXOLOTL_EXTRAS" != "" ] ; then \
-        uv pip install --no-build-isolation -e .[deepspeed,flash-attn,ring-flash-attn,optimizers,ray,$AXOLOTL_EXTRAS] $AXOLOTL_ARGS; \
+        uv pip install --no-build-isolation -e .[ring-flash-attn,optimizers,ray,$AXOLOTL_EXTRAS] $AXOLOTL_ARGS; \
     else \
-        uv pip install --no-build-isolation -e .[deepspeed,flash-attn,ring-flash-attn,optimizers,ray] $AXOLOTL_ARGS; \
+        uv pip install --no-build-isolation -e .[ring-flash-attn,optimizers,ray] $AXOLOTL_ARGS; \
     fi && \
     python scripts/unsloth_install.py | sh && \
     python scripts/cutcrossentropy_install.py | sh && \
diff --git a/docs/installation.qmd b/docs/installation.qmd
index e2dbad6a9..7131c853c 100644
--- a/docs/installation.qmd
+++ b/docs/installation.qmd
@@ -38,7 +38,7 @@ For a quick installation with uv:
 curl -LsSf https://astral.sh/uv/install.sh | sh
 
 # Install axolotl
-uv pip install --no-build-isolation axolotl[flash-attn,deepspeed]
+uv pip install --no-build-isolation axolotl
 ```
 
 ### PyPI Installation {#sec-pypi}
@@ -97,8 +97,9 @@ git clone https://github.com/axolotl-ai-cloud/axolotl.git
 cd axolotl
 # Install uv if not already installed
 curl -LsSf https://astral.sh/uv/install.sh | sh
-source $HOME/.local/bin/env
-uv pip install --no-build-isolation -e '.[flash-attn,deepspeed]'
+# Create and sync environment
+uv venv
+uv sync
 ```
 
 #### Using pip
diff --git a/pyproject.toml b/pyproject.toml
index 6dd72d205..f5a371c2d 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -101,6 +101,8 @@ dependencies = [
     "torchao==0.12.0 ; sys_platform != 'darwin'",
     "bitsandbytes==0.47.0 ; sys_platform != 'darwin'",
     "flash-attn==2.8.3 ; sys_platform == 'linux'",
+    "deepspeed>=0.17.5 ; sys_platform != 'darwin'",
+    "deepspeed-kernels ; sys_platform != 'darwin'",
 ]
 
 [project.optional-dependencies]
@@ -110,12 +112,6 @@ ring-flash-attn = [
     "yunchang==0.6.0",
 ]
 
-# Deep learning frameworks
-deepspeed = [
-    "deepspeed>=0.17.5",
-    "deepspeed-kernels",
-]
-
 mamba-ssm = [
     "mamba-ssm>=2.2.0",
     "causal_conv1d>=1.4.0",