NVTabular Support Matrix

Attention

The support matrix for NVTabular is no longer maintained. See the Merlin Support Matrix page for information about containers that include NVTabular.

We offer the following containers:

  • Merlin Inference: Allows you to deploy NVTabular workflows and HugeCTR or TensorFlow models to the Triton Inference server for production.

  • Merlin Training: Allows you to do preprocessing and feature engineering with NVTabular so that you can train a deep learning recommendation model with HugeCTR.

  • Merlin TensorFlow Training: Allows you to do preprocessing and feature engineering with NVTabular so that you can train a deep learning recommendation model with TensorFlow.

  • Merlin PyTorch Training: Allows you to do preprocessing and feature engineering with NVTabular so that you can train a deep learning recommendation model with PyTorch.

The following tables provide the software and model versions that NVTabular supports per container.


Table 1: Support matrix for the Merlin Inference (merlin-inference) container

Version 21.09

Version 21.11

Version 21.12

Version 22.01

Version 22.02

Version 22.03

DGX

DGX System

  • DGX-1

  • DGX-2

  • DGX A100

  • DGX Station

Operating System

Ubuntu version 20.04

NVIDIA Certified Systems

NVIDIA Driver

NVIDIA Driver version 465.19.01 or later is required. However, if you’re running on Data Center GPUs (formerly Tesla) such as T4, you can use any of the following NVIDIA Driver versions:

  • 418.40 (or later R418)

  • 440.33 (or later R440)

  • 450.51 (or later R450)

  • 460.27 (or later R460)

NOTE: The CUDA Driver Compatibility Package doesn’t support all drivers.

GPU Model

Base Container Image

Container Operating System

Ubuntu version 20.04

Base Container

Triton version 21.07

Triton version 21.10

Triton version 21.11

Triton version 21.12

Triton version 22.02

CUDA

11.6

11.6.55

RMM

21.08

21.10

21.12

21.12

cuDF

21.06

21.08

21.10

21.12

21.12.02

cuDNN

N/A

HugeCTR

3.2

3.2.1

3.3

3.3.1

3.4

3.4.1

NVTabular

0.7

0.7.1

0.8

0.9

0.10

0.11.0

SM

60; 61; 70; 75; 80


Table 2: Support matrix for the Merlin Training (merlin-training) container

Version 21.09

Version 21.11

Version 21.12

Version 22.01

Version 22.02

Version 22.03

DGX

DGX System

  • DGX-1

  • DGX-2

  • DGX A100

  • DGX Station

Operating System

Ubuntu version 20.04

NVIDIA Certified Systems

NVIDIA Driver

NVIDIA Driver version 465.19.01 or later is required. However, if you’re running on Data Center GPUs (formerly Tesla) such as T4, you can use any of the following NVIDIA Driver versions:

  • 418.40 (or later R418)

  • 440.33 (or later R440)

  • 450.51 (or later R450)

  • 460.27 (or later R460)

NOTE: The CUDA Driver Compatibility Package doesn’t support all drivers.

GPU Model

Base Container Image

Container Operating System

Ubuntu version 20.04

Base Container

N/A

CUDA

11.6

11.6.55

RMM

21.08

21.10

21.12

21.12.0a0+31.g0acbd51

cuDF

21.08

21.10

21.12

21.12.00a+293.g0930f712e6

cuDNN

8.3.0

8.3.2

HugeCTR

3.2

3.2.1

3.3

3.3.1

3.4

3.4.1

NVTabular

0.7

0.7.1

0.8

0.9

0.10

0.11.0

SM

60; 61; 70; 75; 80


Table 3: Support matrix for the Merlin TensorFlow Training (merlin-tensorflow-training) container

Version 21.09

Version 21.11

Version 21.12

Version 22.01

Version 22.02

Version 22.03

DGX

DGX System

  • DGX-1

  • DGX-2

  • DGX A100

  • DGX Station

Operating System

Ubuntu version 20.04

NVIDIA Certified Systems

NVIDIA Driver

NVIDIA Driver version 465.19.01 or later is required. However, if you’re running on Data Center GPUs (formerly Tesla) such as T4, you can use any of the following NVIDIA Driver versions:

  • 418.40 (or later R418)

  • 440.33 (or later R440)

  • 450.51 (or later R450)

  • 460.27 (or later R460)

NOTE: The CUDA Driver Compatibility Package doesn’t support all drivers.

GPU Model

Base Container Image

Container Operating System

Ubuntu version 20.04

Base Container

nvcr.io/nvidia/tensorflow:21.07-tf2-py3

*Customized with TensorFlow version 2.4.2

nvcr.io/nvidia/tensorflow:21.10-tf2-py3

nvcr.io/nvidia/tensorflow:21.11-tf2-py3

nvcr.io/nvidia/tensorflow:12.12-tf2-py3

nvcr.io/nvidia/tensorflow:22.02-tf2-py3

CUDA

11.6

11.6.55

RMM

21.08

21.10

21.12

21.12.0a0+31.g0acbd51

cuDF

21.08

21.10

21.12

21.12.0a0+293.g0930f712e6

cuDNN

N/A

HugeCTR

3.2

3.2.1

3.3

3.3.1

3.4

3.4.1

NVTabular

0.7

0.7.1

0.8

0.9

0.10

0.11.0

SM

60; 61; 70; 75; 80


Table 4: Support matrix for the Merlin PyTorch Training (merlin-pytorch-training) container

Version 21.09

Version 21.11

Version 21.12

Version 22.01

Version 22.02

Version 22.03

DGX

DGX System

  • DGX-1

  • DGX-2

  • DGX A100

  • DGX Station

Operating System

Ubuntu version 20.04

NVIDIA Certified Systems

NVIDIA Driver

NVIDIA Driver version 465.19.01 or later is required. However, if you’re running on Data Center GPUs (formerly Tesla) such as T4, you can use any of the following NVIDIA Driver versions:

  • 418.40 (or later R418)

  • 440.33 (or later R440)

  • 450.51 (or later R450)

  • 460.27 (or later R460)

NOTE: The CUDA Driver Compatibility Package doesn’t support all drivers.

GPU Model

Base Container Image

Container Operating System

Ubuntu version 20.04

Base Container

nvcr.io/nvidia/pytorch:21.07-py3

nvcr.io/nvidia/pytorch:21.10-py3

nvcr.io/nvidia/pytorch:21.11-py3

nvcr.io/nvidia/pytorch:21.12-py3

CUDA

11.6

11.6.55

RMM

21.08

21.10

21.12

21.12.0a0+31.g0acbd51

cuDF

21.08

21.10

21.12

21.12.0a0+293.g0930f712e6

cuDNN

N/A

HugeCTR

N/A

NVTabular

0.7

0.7.1

0.8

0.9

0.10

0.11.0

SM

60; 61; 70; 75; 80