From e38b3c340fb52ae4dba01c9513ad2a854b4306cc Mon Sep 17 00:00:00 2001 From: Chuck Tang Date: Mon, 3 Jun 2024 22:04:37 -0700 Subject: [PATCH 1/4] commit change --- Dockerfile | 3 +++ 1 file changed, 3 insertions(+) diff --git a/Dockerfile b/Dockerfile index fd5b8e40ab..98f6224773 100644 --- a/Dockerfile +++ b/Dockerfile @@ -12,6 +12,9 @@ ARG DEP_GROUPS ADD https://raw.githubusercontent.com/mosaicml/llm-foundry/$BRANCH_NAME/setup.py setup.py RUN rm setup.py +# Install TransformerEngine +RUN NVTE_FRAMEWORK=pytorch CMAKE_BUILD_PARALLEL_LEVEL=4 MAX_JOBS=4 pip install git+https://github.com/denera/TransformerEngine.git@databricks/distribute-fp8-weights-fsdp + # Install and uninstall foundry to cache foundry requirements RUN git clone -b $BRANCH_NAME https://github.com/mosaicml/llm-foundry.git RUN pip install --no-cache-dir "./llm-foundry${DEP_GROUPS}" From 3a2e2a0a9aacaf56eb2e5a9877eb530d26279f43 Mon Sep 17 00:00:00 2001 From: Charles Tang Date: Mon, 3 Jun 2024 22:41:46 -0700 Subject: [PATCH 2/4] Update Dockerfile --- Dockerfile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Dockerfile b/Dockerfile index 98f6224773..7bd7c9235a 100644 --- a/Dockerfile +++ b/Dockerfile @@ -13,7 +13,7 @@ ADD https://raw.githubusercontent.com/mosaicml/llm-foundry/$BRANCH_NAME/setup.py RUN rm setup.py # Install TransformerEngine -RUN NVTE_FRAMEWORK=pytorch CMAKE_BUILD_PARALLEL_LEVEL=4 MAX_JOBS=4 pip install git+https://github.com/denera/TransformerEngine.git@databricks/distribute-fp8-weights-fsdp +RUN NVTE_FRAMEWORK=pytorch CMAKE_BUILD_PARALLEL_LEVEL=4 MAX_JOBS=4 pip install git+https://github.com/cli99/TransformerEngine.git@6b21f606f2459d49c2113d69236d68d334edeb4c # Install and uninstall foundry to cache foundry requirements RUN git clone -b $BRANCH_NAME https://github.com/mosaicml/llm-foundry.git From 06ca8fcaa3cbe20f52a992b0d1a9c91b76925a00 Mon Sep 17 00:00:00 2001 From: Charles Tang Date: Tue, 4 Jun 2024 00:15:05 -0700 Subject: [PATCH 3/4] Update Dockerfile --- Dockerfile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Dockerfile b/Dockerfile index 7bd7c9235a..049fd722da 100644 --- a/Dockerfile +++ b/Dockerfile @@ -13,7 +13,7 @@ ADD https://raw.githubusercontent.com/mosaicml/llm-foundry/$BRANCH_NAME/setup.py RUN rm setup.py # Install TransformerEngine -RUN NVTE_FRAMEWORK=pytorch CMAKE_BUILD_PARALLEL_LEVEL=4 MAX_JOBS=4 pip install git+https://github.com/cli99/TransformerEngine.git@6b21f606f2459d49c2113d69236d68d334edeb4c +RUN NVTE_FRAMEWORK=pytorch CMAKE_BUILD_PARALLEL_LEVEL=5 MAX_JOBS=5 pip install git+https://github.com/cli99/TransformerEngine.git@6b21f606f2459d49c2113d69236d68d334edeb4c # Install and uninstall foundry to cache foundry requirements RUN git clone -b $BRANCH_NAME https://github.com/mosaicml/llm-foundry.git From 075b4c1cba45dc456f500f8730f644c74d28d711 Mon Sep 17 00:00:00 2001 From: Charles Tang Date: Tue, 4 Jun 2024 00:41:13 -0700 Subject: [PATCH 4/4] Update Dockerfile --- Dockerfile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Dockerfile b/Dockerfile index 049fd722da..253a5b6cd8 100644 --- a/Dockerfile +++ b/Dockerfile @@ -13,7 +13,7 @@ ADD https://raw.githubusercontent.com/mosaicml/llm-foundry/$BRANCH_NAME/setup.py RUN rm setup.py # Install TransformerEngine -RUN NVTE_FRAMEWORK=pytorch CMAKE_BUILD_PARALLEL_LEVEL=5 MAX_JOBS=5 pip install git+https://github.com/cli99/TransformerEngine.git@6b21f606f2459d49c2113d69236d68d334edeb4c +RUN NVTE_FRAMEWORK=pytorch CMAKE_BUILD_PARALLEL_LEVEL=3 MAX_JOBS=3 pip install git+https://github.com/cli99/TransformerEngine.git@6b21f606f2459d49c2113d69236d68d334edeb4c # Install and uninstall foundry to cache foundry requirements RUN git clone -b $BRANCH_NAME https://github.com/mosaicml/llm-foundry.git