Oracle Cloud Infrastructure Documentation

PyTorch

Learn about PyTorch conda environments.

PyTorch 2.4.0 for GPU on Python 3.10 with CUDA 12.4 (version 1.0)

A description of the PyTorch 2.4.0 for GPU on Python 3.10 with CUDA 12.4 (version 1.0) conda environment.

Released

August 23, 2024

Description

PyTorch is a machine learning library that's used for applications in computer vision and natural language processing. It provides high level features for tensor computing and deep neural networks. Use oracle-ads to speed up your data science workflow with tools that automate common tasks. This environment includes CUDA toolkit and libraries commonly used in training large language models (LLMs).

To get started with the PyTorch environment, open and review the notebook example getting-started.ipynb, using the Launcher.

Python Version

3.10
Slug

pytorch24_p310_gpu_x86_64_v1
Object Storage Path 
oci://service-conda-packs@id19sfcrra6z/service_pack/gpu/PyTorch_2.4_for_GPU_on_Python_3.10/1.0/pytorch24_p310_gpu_x86_64_v1

Top Libraries
  • oracle-ads (v2.11.16)
  • accelerate (v0.33.0)
  • cuda-toolkit (v12.4)
  • deepspeed (v0.14.4)
  • peft (v0.12.0)
  • pytorch (v2.4.0)
  • pytorch-lightning (v2.4.0)
  • tensorboard (v2.12.1)
  • transformers (v4.44.0)
  • torchvision (v0.19.0)
  • langchain (v0.2.13)
  • triton-model-analyzer (v1.42.0)

For a complete list of preinstalled Python libraries, see pytorch24_p310_gpu_x86_64_v1.txt.
PyTorch 2.3.1 for GPU on Python 3.10 with CUDA 12.1 (version 1.0)

A description of the PyTorch 2.3.1 for GPU on Python 3.10 with CUDA 12.1 (version 1.0) conda environment.

Released

August 23, 2024

Description

PyTorch is a machine learning library that's used for applications in computer vision and natural language processing. It provides high level features for tensor computing and deep neural networks. Use oracle-ads to speed up your data science workflow with tools that automate common tasks. This environment includes CUDA toolkit and libraries commonly used in training large language models (LLMs).

To get started with the PyTorch environment, open and review the notebook example getting-started.ipynb, using the Launcher.

Python Version

3.10
Slug

pytorch23_p310_gpu_x86_64_v1
Object Storage Path 
oci://service-conda-packs@id19sfcrra6z/service_pack/gpu/PyTorch_2.3_for_GPU_on_Python_3.10/1.0/pytorch23_p310_gpu_x86_64_v1

Top Libraries
  • oracle-ads (v2.11.16)
  • accelerate (v0.33.0)
  • cuda-toolkit (v12.1)
  • deepspeed (v0.14.4)
  • peft (v0.12.0)
  • pytorch (v2.3.1)
  • pytorch-lightning (v2.3.3)
  • tensorboard (v2.12.1)
  • transformers (v4.44.4)
  • torchvision (v0.18.1)
  • langchain (v0.2.13)
  • triton-model-analyzer (v1.42.0)

For a complete list of preinstalled Python libraries, see pytorch23_p310_gpu_x86_64_v1.txt.
PyTorch 2.1 for GPU on Python 3.9 (version 1.0) with CUDA 11.8

A description of the PyTorch 2.1 for GPU on Python 3.9 with CUDA 11.8 (version 1.0) conda environment.

Released

February 9, 2024

Description

PyTorch is a machine learning library that's used for applications in computer vision and natural language processing. It provides high level features for tensor computing and deep neural networks. Use oracle-ads to speed up your data science workflow with tools that automate common tasks. This environment includes CUDA 11.8 and libraries commonly used in training large language models (LLMs).

To get started with the PyTorch environment, open and review the notebook example getting-started.ipynb, using the Launcher.

Python Version

3.9
Slug

pytorch21_p39_gpu_v1
Object Storage Path 
oci://service-conda-packs@id19sfcrra6z/service_pack/gpu/PyTorch_2.1_for_GPU_on_Python_3.9/1.0/pytorch21_p39_gpu_v1

Top Libraries
  • oracle-ads (v2.10.0)
  • accelerate (v0.26.1)
  • cuda-toolkit (v11.8.0)
  • deepspeed (v0.12.3)
  • peft (v0.7.1)
  • pytorch (v2.1.2)
  • pytorch-lightning (v2.1.2)
  • tensorboard (v2.12.1)
  • transformers (v4.35.2)
  • torchvision (v0.16.1)
  • langchain (v0.1.1)
  • triton-model-analyzer (v1.33.0)

For a complete list of preinstalled Python libraries, see pytorch21_p39_gpu_v1.txt.
PyTorch 2.0 for GPU on Python 3.8 (version 2.0) with CUDA 11.8

A description of the PyTorch 2.0 for GPU on Python 3.8 with CUDA 11.8 (version 2.0) conda environment.

Released

September 12, 2023

Description

PyTorch is a machine learning library that's used for applications in computer vision and natural language processing. It provides high level features for tensor computing and deep neural networks. Use oracle-ads to speed up Data Science workflows with tools that automate common tasks. This environment includes CUDA 11.8 and libraries commonly used in training large language models (LLMs).

To get started with this conda environment, review the Getting Started notebook, using the Launcher.

Python Version

3.8
Slug

pytorch20_p39_gpu_v2
Object Storage Path 
oci://service-conda-packs@id19sfcrra6z/service_pack/gpu/PyTorch_2.0_for_GPU_on_Python_3.9/2.0/pytorch20_p39_gpu_v2

Top Libraries

  • oracle-ads (v2.8.9)

  • accelerate(v0.20.3)

  • cuda-toolkit(v11.8.0)

  • deepspeed (v0.10.0)

  • peft (v0.4.0)

  • pytorch (v2.0.1)

  • pytorch-lightning (v2.0.1)

  • tensorboard (v2.12.1)

  • transformers (v4.24.0)

  • torchvision (v0.15.2)

  • langchain (v0.0.267)

For a complete list of preinstalled Python libraries, see pytorch20_p39_gpu_v2.txt.
PyTorch 2.0 for GPU on Python 3.8 (version 1.0)

A description of the PyTorch 2.0 for GPU on Python 3.8 (version 1.0) conda environment.

Released

July 10, 2022

Description

PyTorch is a machine learning library that's mainly used for applications in computer vision and natural language processing It provides high level features for tensor computing and deep neural networks. Use oracle-ads to speed up your data science work flow with tools that automate common tasks.

To get started with this conda environment, review the Getting Started notebook, using the Launcher.

Python Version

3.8
Slug

pytorch20_p39_gpu_v1
Object Storage Path 
oci://service-conda-packs@id19sfcrra6z/service_pack/gpu/PyTorch_2.0_for_GPU_on_Python_3.9/1.0/pytorch20_p39_gpu_v1

Top Libraries

  • oracle-ads (v2.8.7)

  • pytorch (v2.0.1)

  • pytorch-lightning (v2.0.1)

  • tensorboard (v2.11.0)

  • transformers (v4.24.0)

  • torchvision (v0.15.2)

  • deepspeed (v0.9.4)

  • accelerate (v0.20.3)

For a complete list of preinstalled Python libraries, see pytorch20_p39_gpu_v1.txt.
PyTorch 1.10 for GPU on Python 3.8 (version 1.0)

A description of the PyTorch 1.10 for GPU on Python 3.8 (version 1.0) conda environment.

Released

December 1, 2022

Description

PyTorch is a machine learning library that is mainly used for applications in computer vision and natural language processing It provides high level features for tensor computing and deep neural networks. Use oracle-ads to speed up your data science work flow with tools that automate common tasks.

To get started with this conda environment, review the Getting Started notebook, using the Launcher.

Python Version

3.8
Slug

pytorch110_p38_gpu_v1
Object Storage Path 
oci://service-conda-packs@id19sfcrra6z/service_pack/gpu/PyTorch_1.10_for_GPU_on_Python_3.8/1.0/pytorch110_p38_gpu_v1

Top Libraries

  • oracle-ads (v2.6.8)

  • pandas (v1.4.4)

  • scikit-learn (v1.1.2)

  • pytorch (v1.10.2)

  • pytorch-lightning (v1.7.7)

  • tensorboard (v2.10.1)

  • transformers (v4.18.0)

  • torchvision (v0.11.3)

For a complete list of preinstalled Python libraries, see pytorch110_p38_gpu_v1.txt.