geohot
fb6aaefb9b
save 2 lines
2020-12-29 14:02:50 -05:00
geohot
ea341c84fe
logsoftmax good, div bad
2020-12-29 13:59:39 -05:00
geohot
f18801c7db
simple pool. swimming is very easy now
2020-12-29 13:48:50 -05:00
geohot
8f9232d59b
readmee
2020-12-29 13:40:34 -05:00
geohot
837aaacfbf
Unpad2D on GPU:
2020-12-29 13:16:14 -05:00
geohot
02655c07d5
break maxpool2d on GPU
2020-12-29 13:05:57 -05:00
geohot
061e37de39
touchups
2020-12-29 12:41:21 -05:00
geohot
a2e6562330
fix max op, less lines
2020-12-29 10:47:04 -05:00
Marcel Bischoff and GitHub
dc8fa7999c
Transpose on GPU ( #221 )
...
* 2serious
* load/save
* fixing GPU
* added DEBUG
* needs BatchNorm or doesn't learn anything
* old file not needed
* added conv biases
* added extra/training.py and checkpoint
* assert in test only
* save
* padding
* num_classes
* checkpoint
* checkpoints for padding
* training was broken
* merge
* rotation augmentation
* more aug
* needs testing
* streamline augment, augment is fast thus bicubic
* tidying up
* transformer eval
* axis=-1
* transpose
* test for permutation using torch.movedims
* another test
* line
2020-12-29 10:40:11 -05:00
geohot
36579f66bf
max op
2020-12-28 23:54:52 -05:00
geohot
bcb3ceeca3
set training in functions
2020-12-28 22:45:46 -05:00
geohot
51bf164b72
dropout, training
2020-12-28 22:12:23 -05:00
geohot
7b8fee038d
it works! forgot the sqrt
2020-12-28 16:23:52 -05:00
geohot
1faf05ef67
ahh, it's better if i don't train the embedding
2020-12-28 16:07:02 -05:00
geohot
c3832e1bde
hmm, fix layernorm to not be batchnorm and it breaks
2020-12-28 13:06:21 -05:00
geohot
2e89e75dcb
layernorm fixes transformer instability
2020-12-28 12:58:15 -05:00
geohot
628d21f899
doc touchup
2020-12-28 10:45:26 -05:00
geohot
fafece9db7
avgpool2d is a second class op
2020-12-28 10:41:59 -05:00
geohot
593233b668
log and exp are first class ops
2020-12-28 10:00:30 -05:00
Marcel Bischoff and GitHub
ffff98db78
Evaluation in Transformers ( #218 )
...
* 2serious
* load/save
* fixing GPU
* added DEBUG
* needs BatchNorm or doesn't learn anything
* old file not needed
* added conv biases
* added extra/training.py and checkpoint
* assert in test only
* save
* padding
* num_classes
* checkpoint
* checkpoints for padding
* training was broken
* merge
* rotation augmentation
* more aug
* needs testing
* streamline augment, augment is fast thus bicubic
* tidying up
* transformer eval
2020-12-28 09:24:51 -05:00
geohot
65b07d2f4f
fix onehot embed
2020-12-27 18:50:38 -05:00
geohot
d864e1c71a
transformer is training
2020-12-27 18:46:32 -05:00
geohot
a361ef6861
fixup training loop
2020-12-27 18:35:56 -05:00
geohot
f15bec6dbc
make multidot work on CPU
2020-12-27 17:25:37 -05:00
geohot
131e04c90c
cpu only decorator
2020-12-27 17:18:55 -05:00
geohot
2f1b2c0a3b
add transpose, start on transformer
2020-12-27 16:59:12 -05:00
d379502c04
Cleaning ( #211 )
...
* Cleaned
* Brought the lines into line
Co-authored-by: gamwe6 <[email protected]
2020-12-27 09:58:51 -05:00
geohot
8a335f03ad
clock speed 32x32
2020-12-22 18:18:52 -05:00
geohot
aae2e35208
benchmarking 512x512 GEMM
2020-12-22 18:01:36 -05:00
geohot
bd18e03138
conv from weights works
2020-12-22 17:42:17 -05:00
geohot
b3cf53e39b
more docs
2020-12-22 17:14:38 -05:00
geohot
4065eae0fb
docs for tensor stride
2020-12-22 17:06:36 -05:00
geohot
6fb127d5c7
l2 cache note
2020-12-22 16:48:19 -05:00
geohot
78a06a1285
more readme
2020-12-22 16:23:08 -05:00
geohot
0ab951f21c
better readme
2020-12-22 15:57:33 -05:00
geohot
6ca449afd2
sum works
2020-12-22 12:53:20 -05:00
geohot
ebc7f8305c
3x3 gemm in conv
2020-12-22 12:00:44 -05:00
geohot
ff52f832dc
conv hwx
2020-12-22 11:49:54 -05:00
geohot
5b93d9c03a
aneregs
2020-12-22 10:41:12 -05:00
geohot
ccfd03e401
parse ane struct
2020-12-22 10:17:06 -05:00
Nicklas Boman and GitHub
06f359baa3
issue-193 - Move torch loader out of efficientnet code ( #213 )
2020-12-22 00:19:16 -05:00
geohot
3abda29648
add debug to ane lib
2020-12-21 18:49:33 -05:00
geohot
c48917fdf0
ane reg notes
2020-12-21 15:26:52 -05:00
geohot
451e944648
aneregs from debug
2020-12-21 13:37:02 -05:00
geohot
16f0cb7d80
fix sign python
2020-12-20 15:44:10 -05:00
geohot
e03759f01d
channels concat
2020-12-19 09:43:21 -08:00
geohot
73ae6d7821
working on gemm
2020-12-19 09:30:29 -08:00
Liam and GitHub
075cf4bc02
Update examples to new API ( #205 )
2020-12-19 09:04:12 -08:00
geohot
3c9150a9f4
lil note
2020-12-18 09:54:20 -08:00
geohot
be41e73ad9
ane docs continue
2020-12-18 09:49:13 -08:00
geohot
