Create and Add Two Tensors

We will review a simple example that demonstrates how to create two tensors and add them together using TT-NN, a high-level Python API designed for developers to run models like LLaMA, Mistral, Stable Diffusion, and more on Tenstorrent devices.

Lets create the example file, ttnn_add_tensors.py

Import the necessary libraries

import ttnn
from loguru import logger

Open Tenstorrent device

Create necessary device on which we will run our program.

# Open Tenstorrent device
device = ttnn.open_device(device_id=0)

Tensor Creation

Create two TT-NN tensors, and initialize them with values 1 and 2 respectively. The preferred shape of the tensors is (32, 32) which will match the hardware’s tile size.

# Create two TT-NN tensors with TILE_LAYOUT
     tt_tensor1 = ttnn.full(
         shape=(32, 32),
         fill_value=1.0,
         dtype=ttnn.float32,
         layout=ttnn.TILE_LAYOUT,
         device=device,
     )
     tt_tensor2 = ttnn.full(
         shape=(32, 32),
         fill_value=2.0,
         dtype=ttnn.float32,
         layout=ttnn.TILE_LAYOUT,
         device=device,
     )

Perform the addition operation and convert back

Now we can perform the addition operation on the two TT-NN tensors and log out the result.

# Perform eltwise addition on the device
tt_result = ttnn.add(tt_tensor1, tt_tensor2)

# Log output tensor
logger.info("Output tensor:")
logger.info(tt_result)

Full example and output

Lets put everything together in a complete example that can be run directly. This example will open a Tenstorrent device, create two tensors, perform the addition, and log the output tensor.

Source Code

# SPDX-FileCopyrightText: © 2025 Tenstorrent AI ULC
# SPDX-License-Identifier: Apache-2.0

import ttnn
from loguru import logger


def main():
    # Open Tenstorrent device
    device = ttnn.open_device(device_id=0)

    try:
        # Create two TT-NN tensors with TILE_LAYOUT
        tt_tensor1 = ttnn.full(
            shape=(32, 32),
            fill_value=1.0,
            dtype=ttnn.float32,
            layout=ttnn.TILE_LAYOUT,
            device=device,
        )
        tt_tensor2 = ttnn.full(
            shape=(32, 32),
            fill_value=2.0,
            dtype=ttnn.float32,
            layout=ttnn.TILE_LAYOUT,
            device=device,
        )

        # Log input tensors
        logger.info("Input tensors:")
        logger.info(tt_tensor1)
        logger.info(tt_tensor2)

        # Perform eltwise addition on the device
        tt_result = ttnn.add(tt_tensor1, tt_tensor2)

        # Log output tensor
        logger.info("Output tensor:")
        logger.info(tt_result)

    finally:
        # Close Tenstorrent device
        ttnn.close_device(device)


if __name__ == "__main__":
    main()

Running this script will output the input tensors and the result of their addition, which should be a tensor filled with 3s. As shown below

$ python3 $TT_METAL_HOME/ttnn/tutorials/basic_python/ttnn_add_tensors.py
2025-06-23 09:36:58.211 | INFO     | __main__:main:29 - Input tensors:
2025-06-23 09:36:58.211 | INFO     | __main__:main:30 - ttnn.Tensor([[ 1.00000,  1.00000,  ...,  1.00000,  1.00000],
            [ 1.00000,  1.00000,  ...,  1.00000,  1.00000],
            ...,
            [ 1.00000,  1.00000,  ...,  1.00000,  1.00000],
            [ 1.00000,  1.00000,  ...,  1.00000,  1.00000]], shape=Shape([32, 32]), dtype=DataType::FLOAT32, layout=Layout::TILE)
2025-06-23 09:36:58.211 | INFO     | __main__:main:31 - ttnn.Tensor([[ 2.00000,  2.00000,  ...,  2.00000,  2.00000],
            [ 2.00000,  2.00000,  ...,  2.00000,  2.00000],
            ...,
            [ 2.00000,  2.00000,  ...,  2.00000,  2.00000],
            [ 2.00000,  2.00000,  ...,  2.00000,  2.00000]], shape=Shape([32, 32]), dtype=DataType::FLOAT32, layout=Layout::TILE)
2025-06-23 09:37:00.524 | INFO     | __main__:main:37 - Output tensor:
2025-06-23 09:37:00.525 | INFO     | __main__:main:38 - ttnn.Tensor([[ 3.00000,  3.00000,  ...,  3.00000,  3.00000],
            [ 3.00000,  3.00000,  ...,  3.00000,  3.00000],
            ...,
            [ 3.00000,  3.00000,  ...,  3.00000,  3.00000],
            [ 3.00000,  3.00000,  ...,  3.00000,  3.00000]], shape=Shape([32, 32]), dtype=DataType::FLOAT32, layout=Layout::TILE)