NeMo Framework Megatron Backend

NVIDIA

Container

NVIDIA

NeMo Framework Megatron Backend

NVIDIA NeMo™ framework Megatron backend supports pre-training, post-training, and reinforcement learning of LLMs and multi-modal generative AI models with state-of-the-art data processing, model training techniques, and flexible deployment options.

NVIDIA AI Enterprise Supported

Layer	Label		Created
ab651b1292febca123b20f8bd1ec93b18b02a952d511ba2551bf948dcb1f8a65	CONFIG	`Entrypoint /opt/nvidia/nvidia_entrypoint.sh; WorkingDir /workspace; ExposedPorts 6006/tcp, 8888/tcp`	12/19/2024 6:06 AM UTC
1946beac9195b9dc07892540d3c1913f961844252e7da882330bc57193ea7b47	ARG	`NVIDIA_BUILD_REF`	12/19/2024 6:06 AM UTC
bbfa991665d7a614cc40df3bac17c11955b3910abbe4e8e29f3abece16056551	LABEL	`com.nvidia.build.id=114410972`	12/19/2024 6:06 AM UTC
5109555d7ac581e328a76236c86993241d4da995f90012890a22b9d0e618693a	ENV	`NVIDIA_BUILD_ID=114410972`	12/19/2024 6:06 AM UTC
9e8d444e00f3d73c9652227c447366d890bb8d5e772ff04f448b1d121e8becd0	ARG	`NVIDIA_BUILD_ID`	12/19/2024 6:06 AM UTC
a7b0177c7da0e05db30967e3a98e64cab6b0edf0fc9ef9b8af069a85e16ebb8a	RUN	NVIDIA_BIGNLP_VERSION= TRT_VER=10.5.0.18 CUDA_VER=12.6 CUDNN_VER=9.5.0.50-1 NCCL_VER= CUBLAS_VER=12.6.3.3-1 NVRTC_VER=12.6.77-1 REINSTALL_APEX=False APEX_COMMIT= MCORE_COMMIT=core_r0.10.0 SRC_DIR=/src/tensorrt_llm CPP_BUILD_DIR=/src/tensorrt_llm/cpp/build TARGET_ARCH=x86 INSTALL_VLLM=True ALIGNER_COMMIT=dev LAUNCHER_COMMIT= DATA_CURATOR_COMMIT=r0.6.0 INSTALL_DATA_CURATOR=True MODELOPT_VERSION=0.19.0 NEMO_COMMIT=r2.1.0 RESIL_COMMIT=97aad77609d2e25ed38ac5c99f0c13f93c48464e NEMO_RUN_COMMIT= PRE_COMPILE_QUANT_KERNELS=True chmod -R a+w /workspace	12/19/2024 6:06 AM UTC
f28be5c0e70cd1120a48508763a8d3d8bd23403f33b388c0713502762aedb945	WORKDIR	`/workspace`	12/19/2024 6:06 AM UTC
805f13dff07a6879a1d0b6aca3ee84ec17037a667629c5716154d1fd4cf0a9c1	RUN	NVIDIA_BIGNLP_VERSION= TRT_VER=10.5.0.18 CUDA_VER=12.6 CUDNN_VER=9.5.0.50-1 NCCL_VER= CUBLAS_VER=12.6.3.3-1 NVRTC_VER=12.6.77-1 REINSTALL_APEX=False APEX_COMMIT= MCORE_COMMIT=core_r0.10.0 SRC_DIR=/src/tensorrt_llm CPP_BUILD_DIR=/src/tensorrt_llm/cpp/build TARGET_ARCH=x86 INSTALL_VLLM=True ALIGNER_COMMIT=dev LAUNCHER_COMMIT= DATA_CURATOR_COMMIT=r0.6.0 INSTALL_DATA_CURATOR=True MODELOPT_VERSION=0.19.0 NEMO_COMMIT=r2.1.0 RESIL_COMMIT=97aad77609d2e25ed38ac5c99f0c13f93c48464e NEMO_RUN_COMMIT= PRE_COMPILE_QUANT_KERNELS=True if [ $PRE_COMPILE_QUANT_KERNELS = "True" ]; then python -c "import modelopt.torch.quantization.extensions as ext; ext.precompile()"; fi	12/19/2024 6:06 AM UTC
026ff1534d96b654266ddf3ffb222f9cab798dc05b589d71b588e34e296673c3	ARG	`PRE_COMPILE_QUANT_KERNELS=True`	12/19/2024 5:58 AM UTC
c76188812a8185925e05fbd66c27d43af7eadf70a14bc56b6ef11333829ce258	RUN	NVIDIA_BIGNLP_VERSION= TRT_VER=10.5.0.18 CUDA_VER=12.6 CUDNN_VER=9.5.0.50-1 NCCL_VER= CUBLAS_VER=12.6.3.3-1 NVRTC_VER=12.6.77-1 REINSTALL_APEX=False APEX_COMMIT= MCORE_COMMIT=core_r0.10.0 SRC_DIR=/src/tensorrt_llm CPP_BUILD_DIR=/src/tensorrt_llm/cpp/build TARGET_ARCH=x86 INSTALL_VLLM=True ALIGNER_COMMIT=dev LAUNCHER_COMMIT= DATA_CURATOR_COMMIT=r0.6.0 INSTALL_DATA_CURATOR=True MODELOPT_VERSION=0.19.0 NEMO_COMMIT=r2.1.0 RESIL_COMMIT=97aad77609d2e25ed38ac5c99f0c13f93c48464e NEMO_RUN_COMMIT= sed -i '/^LayerId/s/^/#/' /usr/local/lib/python3.10/dist-packages/cv2/typing/__init__.py	12/19/2024 5:58 AM UTC