George Hotz
b4bf6a7dea
switch backward to use gradient [pr] (#8235)
* switch backward to use gradient [pr]
* set device correctly, dedup
* why does that fail?
* add noop cast
* simple backward
* fix beautiful_mnist
* touchups
* set in compute_gradient
* uop_count
* uop_count was wrong
* collections
* no note
* skip that test
* update sched kernel counts
* train mnist is 65
* fix metadata and gc
* fixes
* materialize_grads
* no pathlib stuff
* add contiguous_backward, fix bugs
* add some realize
* fix multi
2025-01-26 09:12:16 +09:00
..
2023-12-05 16:17:57 -08:00
2025-01-10 18:21:01 -05:00
2024-12-12 16:32:36 -05:00
2024-05-05 14:19:01 -04:00
2025-01-21 09:57:47 -08:00
2024-08-10 11:17:56 -07:00
2024-12-16 14:28:10 -05:00
2025-01-26 09:12:16 +09:00
2025-01-20 14:56:27 -05:00
2024-12-27 17:27:14 -05:00
2024-11-16 20:56:56 +08:00
2024-11-21 12:18:05 +08:00
2025-01-14 16:03:17 -05:00
2024-12-22 05:21:03 -05:00
2025-01-19 17:03:12 -08:00
2024-10-25 17:05:09 +07:00
2025-01-19 17:03:12 -08:00
2024-12-05 11:40:59 +08:00
2024-12-04 17:03:19 +08:00
2025-01-21 09:57:47 -08:00
2024-11-18 14:12:03 -05:00
2024-10-31 18:41:22 -04:00
2024-10-29 16:43:04 +08:00
2025-01-19 17:03:12 -08:00
2025-01-01 10:21:59 -05:00
2025-01-13 06:24:11 -05:00
2025-01-08 22:11:24 -05:00
2025-01-25 19:41:57 +02:00