Float accuracy visualization I made a float accuracy…
0 reactions
Float accuracy visualization
I made a float accuracy visualization showing difference between double (64-bit) and single (32-bit), half (16-bit), and fp8 (8-bit).
I haven't seen it done in this format and thought it looks interesting!
Website: https://spievniev.github.io/FloatMap
Repo: https://github.com/spievniev/FloatMap
#technology
I made a float accuracy visualization showing difference between double (64-bit) and single (32-bit), half (16-bit), and fp8 (8-bit).
I haven't seen it done in this format and thought it looks interesting!
Website: https://spievniev.github.io/FloatMap
Repo: https://github.com/spievniev/FloatMap
#technology
| 4,000 mlx | |
| 6 | |
| 0 |