It would be interesting to revisit this review for the forthcoming Strix Halo. Particularly, for ML tasks on relative large memory sizes which are prohibitively

It would be interesting to revisit this review for the forthcoming Strix Halo. Particularly, for ML tasks on relative large memory sizes which are prohibitively expensive for GPUs.