George Hotz
6b0aa2a902
sorry about the line count, this is a good optimization
2022-07-03 17:11:13 -07:00
George Hotz
748618530b
tests will run at okay speed now?
2022-07-03 16:41:52 -07:00
George Hotz
c3d13893f9
add SHUFFLE_MOVEMENT_OPS, exactly 1000 lines
2022-07-03 16:30:42 -07:00
George Hotz
e6e43e820e
should fix tests
2022-07-03 16:06:11 -07:00
George Hotz
71a812fbf2
elementwise_ops
2022-07-03 15:29:38 -07:00
George Hotz
d7aad46758
test lazy also, make TestMNIST faster
2022-07-03 15:19:19 -07:00
Nicklas Boman
64d986bc8b
add mypy to ci testing ( #353 )
2022-07-03 15:11:35 -07:00
George Hotz
57ebce8d67
first LazyBuffer optimizations
2022-07-03 15:09:16 -07:00
George Hotz
a1a20891ef
more types
2022-07-03 14:03:34 -07:00
George Hotz
99b287ed87
typechecks
2022-07-03 13:54:30 -07:00
George Hotz
cdf2be74f9
add neg
2022-07-03 13:04:58 -07:00
George Hotz
72a9ff7011
remove numpy usage
2022-07-03 12:58:51 -07:00
George Hotz
745e36fda5
mlops cleanup
2022-07-03 12:41:05 -07:00
George Hotz
93c378dffc
add test for slice_one
2022-07-03 12:14:20 -07:00
George Hotz
d10dd175f4
fix len 0 shapes in getitem
2022-07-03 12:12:02 -07:00
George Hotz
1b1c82fac7
print underlying buffer if it's realized
2022-07-03 11:52:58 -07:00
George Hotz
df16b455a7
make lazy the default ( #352 )
...
* make lazy the default
* always float32
* while the lazy framework should be default, lazyness itself shouldn't be (for now)
* bugfixes
* remove the need for the ops class
* fxn_for_op
* hmm, my contiguous asserts went away
* move small shape thing
* refactor reduce
* remove the weird unused new functions
* only that install works
* thats broken
* unused imports, should be good if it passes
2022-07-03 11:40:27 -07:00
George Hotz
bbfdd28a6d
flops counter was dumb
2022-07-03 07:41:52 -07:00
George Hotz
c7a580daa9
flip div order
2022-07-02 23:37:22 -07:00
George Hotz
0d82cfd587
huh, torch 1.12 broke it. remove unused requirements.txt and pin torch 1.11
2022-07-02 23:07:59 -07:00
George Hotz
e822aae9ec
reorg opts, nicer graph
2022-07-02 22:29:09 -07:00
George Hotz
f9a8412b68
make contiguous ops yellow
2022-07-02 17:54:04 -07:00
George Hotz
207b9e1df3
padding is now a param to conv2d
2022-07-02 17:11:12 -07:00
George Hotz
cde137d163
simple shapetracker tests
2022-07-02 16:02:15 -07:00
George Hotz
368c0ce2f6
NUM=-2 for ants
2022-07-02 15:47:10 -07:00
George Hotz
7276f8d6bf
improve constant folding, detach before moving tensor
2022-07-02 15:29:40 -07:00
George Hotz
0cb99d72e9
NUM=-1 is a small efficientnet for small people
2022-07-02 15:11:51 -07:00
George Hotz
8cf1aed0f4
don't track_running_stats, parameters must require_grad
2022-07-02 14:38:45 -07:00
George Hotz
07b438aa8b
move that to resolve time
2022-07-02 14:26:13 -07:00
George Hotz
dbf4aa09db
assert and tuple
2022-06-27 09:19:54 -07:00
George Hotz
37a6c0ef59
create with new ShapeTracker
2022-06-27 09:07:45 -07:00
George Hotz
e55a9833fb
a little more readable
2022-06-27 08:54:04 -07:00
George Hotz
67ff6b52fd
move padding to convs in enet
2022-06-26 23:14:31 -07:00
George Hotz
04f521a963
err, float32
2022-06-26 23:05:04 -07:00
George Hotz
8540d1f289
track membw
2022-06-26 23:03:53 -07:00
George Hotz
3a414d7f50
cleanup, add flops tracking
2022-06-26 22:43:39 -07:00
George Hotz
a699f7cb0b
debug cleanups
2022-06-26 21:58:44 -07:00
George Hotz
15a16b98e6
remove get_root
2022-06-26 21:18:02 -07:00
George Hotz
e3c2579537
flip stride to match canonical
2022-06-26 19:19:53 -07:00
George Hotz
53ab09de79
remove the SLICE on conv dw
2022-06-26 19:09:36 -07:00
George Hotz
149581b0b2
Cdx without SLICE
2022-06-26 18:51:53 -07:00
George Hotz
a04813ffe3
1 line less in cpu, fix torch tests
2022-06-26 18:11:53 -07:00
George Hotz
dffde3de5a
support both asymmetric and negative padding
2022-06-26 17:59:25 -07:00
George Hotz
49c954b389
comments
2022-06-26 17:20:25 -07:00
George Hotz
8c483fbdc9
maxpool lazy fix
2022-06-26 17:07:03 -07:00
George Hotz
f607f18006
fix backward
2022-06-25 00:00:53 -07:00
George Hotz
ec30f0402f
improve benchmark_train_efficientnet
2022-06-24 23:46:38 -07:00
George Hotz
3a147137ee
CL_DEVICE option
2022-06-24 23:22:10 -07:00
George Hotz
d748353ce5
err, okay, a bit more off
2022-06-24 22:44:57 -07:00
George Hotz
bdde95f16e
CACHE_LAZYBUFFERS options + benchmark. only a couple x from torch
2022-06-24 22:33:53 -07:00