granite-20b-code-base-8k
Published checkpoint tensor payload: 37.4 GiB. This includes stored copies and buffers; it is separate from the runtime estimates below.
Weight-only estimates: 16-bit: 37.4 GiB; 8-bit: 20.6 GiB; 4-bit: 10.3 GiB; Cache geometry is unavailable. These are not complete GPU requirements.