Hi Matteo,

I had a quick look at your document; I liked it and I think it's a good starting point. One thing I see is missing in the doc is the testing strategy for sparse tensors: we'll need to think carefully what new tests need to be implemented for these new tensors.

Some comments about your doc:

1. Any new methods like is_sparse() should be added to ITensorInfo rather than ITensor 2. We already have data_layout() method which returns NHWC/NCHW. See https://github.com/ARM-software/ComputeLibrary/blob/main/arm_compute/core/IT... . For sparse tensors I'd use a different name for the new method describing the sparse format, something like sparse_layout()/format()

Hope this helps

________________________________ From: Matteo Nicoli matteo.nicoli001@gmail.com Sent: 02 July 2025 11:30 To: Pablo Tello Pablo.Tello@arm.com Cc: Gunes Bayir Gunes.Bayir@arm.com; acl-dev@lists.linaro.org acl-dev@lists.linaro.org Subject: Re: A draft for sparse tensors design

Thank you Pablo, look forward to hearing from you.

Matteo

On Jul 2, 2025, at 10:54 AM, Pablo Tello Pablo.Tello@arm.com wrote:

Hi Matteo,

Thanks for your contribution last week and sharing this design document with us. We'll have a look at it and get back to you with some feedback.

Sparse tensors support aligns well with the roadmap for ACL and we would be happy to review and take in your patch implementing this feature.

Hope this helps ________________________________ From: Matteo Nicoli <matteo.nicoli001@gmail.commailto:matteo.nicoli001@gmail.com> Sent: 30 June 2025 17:33 To: Gunes Bayir <Gunes.Bayir@arm.commailto:Gunes.Bayir@arm.com>; Pablo Tello <Pablo.Tello@arm.commailto:Pablo.Tello@arm.com> Cc: acl-dev@lists.linaro.orgmailto:acl-dev@lists.linaro.org <acl-dev@lists.linaro.orgmailto:acl-dev@lists.linaro.org> Subject: A draft for sparse tensors design

Dear Gunes and Pablo,

I noticed in the comments of the issue #1084https://github.com/ARM-software/ComputeLibrary/issues/1084 that someone discussed the introduction of sparse tensors a while back. Although the last comment in the thread states, “There is currently no planned work to implement this feature. I’ll discuss it again with the team,” I thought the issue hadn’t been closed, indicating some interest in it. On Saturday, I began thinking on how this feature could be integrated into the codebase. I’ve drafted a design for these classes, which I’d like to share with you.

I understand that this feature may not be a priority for your team at the moment. However, if it doesn’t contradict your roadmap, once issue #1169https://github.com/ARM-software/ComputeLibrary/issues/1169 is resolved, I’ll be able to begin implementing it myself on an experimental basis.

Best regards, Matteo

IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.

IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.