geohot
5486135f2d
readme
2020-11-07 11:41:27 -08:00
geohot
9b8fefe4e3
fix broadcast bugggg
2020-11-07 11:38:56 -08:00
geohot
b1ca4dd327
who loves speeeeed
2020-11-07 11:36:52 -08:00
geohot
e6c8321e5b
enet is running on GPU
2020-11-07 11:33:46 -08:00
geohot
30442a086a
some broadcasting, pool test is fail
2020-11-07 11:29:42 -08:00
Ryan Neph and GitHub
7c630ca355
GPU {Avg,Max}Pool2D forward pass ( #66 )
2020-11-07 11:13:24 -08:00
geohot
40c538c299
fix reshape
2020-11-07 11:10:29 -08:00
geohot
19c2950310
refactor ops to prepare for broadcasting
2020-11-07 11:02:45 -08:00
geohot
94d44c97bf
add pad2d on GPU
2020-11-07 10:46:36 -08:00
geohot
fbff6ab2e5
fix strided convs, GPU env var for enet
2020-11-07 10:26:37 -08:00
geohot
ec03eb44bd
tinygrad does forward pass convs on GPU
2020-11-07 10:15:56 -08:00
Manjeet Singh and GitHub
23c39d9f52
fetch() 'ran out of input' fix ( #65 )
...
If requests.get fails during a download, the incomplete file is passed
without a size check. This causes errors during image/model loading.
Not sure if this is a problem on all systems.
2020-11-07 09:45:10 -08:00
geohot
bc7758cc5b
getting convs to work on gpu
2020-11-07 09:17:57 -08:00
geohot
3302286e68
yayay test_sgd_gpu passes
2020-11-07 08:48:17 -08:00
Ryan Neph and GitHub
98d1a4f740
GPU Selection works when len(answers)>ngpus ( #64 )
...
Fixes TypeError: sequence item 0: expected str instance, int found
triggered by https://github.com/inducer/pyopencl/blob/v2020.2.2/pyopencl/__init__.py#L1500
in pyopencl v2020.2.2
2020-11-07 08:06:59 -08:00
geohot
ce0c80e1bd
ugh, corrupt downloads. when did this break?
2020-11-07 08:04:43 -08:00
geohot
9332fe1989
weee, logsoftmax on GPU. very slow
2020-11-07 07:59:37 -08:00
geohot
38e112cccd
logsoftmax test
2020-11-07 07:26:53 -08:00
geohot
237bbe3625
sigmoid on GPU
2020-11-07 07:23:46 -08:00
Rakib Fiha and GitHub
f40dbd791c
Use --upgrade since its in active dev ( #63 )
2020-11-07 07:15:05 -08:00
Rene Delgado and GitHub
cd54697fd8
fix gpu sum forward ( #61 )
...
* ignore venv
* add sum test
* fix sum forward
2020-11-05 21:59:16 -08:00
gallanoe and GitHub
aea1069f63
Div on CPU ( #58 )
...
* Added Div on CPU
* Removed eps. value
* Fixed tabs
2020-11-05 21:58:37 -08:00
NeuralLink and GitHub
cc605da36d
Stable Sigmoid op ( #59 )
...
* 🔨 Added stable sigmoid
* ✅ added sigmoid test
* 🔧 suppressed overflow warning
* 🔧 clean up
2020-11-05 21:57:50 -08:00
George Hotz
cc852efa64
make OpenCL work elsewhere
2020-11-05 07:59:55 -08:00
Ryan Neph and GitHub
22a5f9975d
pow+div on GPU ( #57 )
2020-11-05 07:49:45 -08:00
b16fadc5c6
Speed up sum backward pass ( #54 )
...
Co-authored-by: holonomicjl <[email protected]
2020-11-04 19:55:45 -08:00
George Hotz
0be3d2cdba
install pyopencl
2020-11-04 11:22:11 -08:00
George Hotz
3efb4f4df4
chicken.jpg
2020-11-04 11:20:22 -08:00
Marcel Bischoff and GitHub
c69582744d
Update efficientnet.py ( #53 )
...
from https://github.com/lukemelas/EfficientNet-PyTorch/blob/master/efficientnet_pytorch/utils.py
```
blocks_args = [
'r1_k3_s11_e1_i32_o16_se0.25',
'r2_k3_s22_e6_i16_o24_se0.25',
'r2_k5_s22_e6_i24_o40_se0.25',
'r3_k3_s22_e6_i40_o80_se0.25',
'r3_k5_s11_e6_i80_o112_se0.25',
'r4_k5_s22_e6_i112_o192_se0.25',
'r1_k3_s11_e6_i192_o320_se0.25',
]
```
now it's a persian cat.
2020-11-04 09:36:27 -08:00
Marcel Bischoff and GitHub
c40862241d
Revert "choose aspect ration dependent on landscape or portrait ( #48 )" ( #51 )
...
This reverts commit 2808411421
2020-11-02 19:31:07 -08:00
Anders Aaen Springborg and GitHub
2808411421
choose aspect ration dependent on landscape or portrait ( #48 )
2020-11-02 18:10:22 -08:00
Marcel Bischoff and GitHub
d24363f421
Update setup.py ( #49 )
...
I think `:=` in tinygrad/test/test_mnist.py actually needs 3.8
2020-11-02 18:09:31 -08:00
geohot
940e14c6ca
more readme
2020-11-02 08:33:48 -08:00
geohot
1e6bbdf4f8
readme updates
2020-11-02 08:30:43 -08:00
geohot
f178d23ff3
gpu relu is good
2020-11-02 08:25:32 -08:00
geohot
8766346187
fix gpu sub
2020-11-02 08:18:58 -08:00
geohot
231c1134bd
cute trick for GPU test
2020-11-02 08:17:17 -08:00
geohot
6a7896f3e9
opsgpu cl program caching
2020-11-02 07:46:26 -08:00
geohot
2e7f16bf3f
the power of cheating
2020-11-02 07:42:11 -08:00
geohot
fc358a07ad
print sloccount
2020-11-02 07:38:13 -08:00
geohot
3da64acd6b
the power of small
2020-11-02 07:36:34 -08:00
geohot
82fc842b40
in_place_op
2020-11-02 07:26:13 -08:00
geohot
1d793b8571
gpu SGD runs now
2020-11-02 07:19:12 -08:00
geohot
621c61b30d
sgd now operates on Tensors
2020-11-02 07:14:54 -08:00
geohot
355402504e
ops work
2020-11-02 07:03:23 -08:00
af5a4e0f5a
alternative to einsum which is slow in extreme cases ( #44 )
...
* streamlined numerical_jacobian
* Got rid of the g loop in Conv2D.forward
* ereased stupid line
* nothing
* no loops in Conv2D forward
* Conv2D backprop improved
* stupid things in examples
* alternative to einsum
* Conv2D backward einsum alternative
* tidying up
* tidied up
* no ravel
* got rid of print
* Update efficientnet.py
* Update efficientnet.py
* Update efficientnet.py
* only tensordot
* 255.0
* whitespace
* aspect ratio error in efficientnet
* noprint
Co-authored-by: Marcel Bischoff <[email protected]
2020-11-02 06:36:07 -08:00
geohot
5201a8e89f
matmul on GPU
2020-11-01 08:54:20 -08:00
geohot
41e7d59aed
test dot
2020-11-01 07:51:35 -08:00
geohot
1f544d6ece
test mnist on GPU
2020-11-01 07:46:17 -08:00
geohot
499604d69b
convert tensor to gpu
2020-11-01 07:34:50 -08:00
geohot
