Update ort CIs (slow, gpu, train) #2024
echarlaix
merged
73 commits into
main
from
enable-ort-gpu-tests
update ort CIs
17bc1719
fix train ci
fbaa980b
fix gpu ci
90aa85d8
gpus all
87c9f3e7
devel
0c1c6bd7
enable trt
430260e0
fix
00e51c7b
fix
3fc54869
fix
80442326
test
2fd4d478
rename
1f322fc7
change instance
6f7c599d
test
806faca7
use available
3eecee6b
Merge branch 'main' into enable-ort-gpu-tests
ab62319a
Merge branch 'main' into enable-ort-gpu-tests
1b7e6528
update
cebe6bfe
shorter labels as well
d0f62b05
add onnxruntime-traning
d001b9b4
Merge branch 'main' into enable-ort-gpu-tests
d2716374
fix onnxruntime package checking
a318c0a2
Merge branch 'enable-ort-gpu-tests' of https://github.com/huggingface…
75976926
fix typo
a6b3a8ec
fix typo
a5c76c42
remove torch version
745ad8dd
fix trainer
bb48c4dc
fixed trt ep by using trt docker image (the only way to make sure eve…
0518dfdb
latest trt version
9635ec41
remove pkv speedup timing since never used
cb9cb7f8
trust remote code for training datasets
eb25460e
remove rocm from diffusers tests
0a7a23d1
move ort training tests to onnxruntime-training
64e9c86a
fix ort training
bbed6bcb
fix
13342008
style
84bf7eef
always assert closenes and not equality
be10d268
fixed perceiver
7ba72a60
fixed missing position ids when attn mask is given
eceba5b8
remove num_labels from output shapes as it's not a dynamic axis
9150e059
raise error on missing mandatory inputs
198ce060
added atol and rtol as part of the ORTModelTestMixin class
930103f4
fix segformer image segmentation
49cfdc02
style
5b8efd40
fix vision encoder io binding
941484a6
hot fix io binding, remove its dependency to the order of inputs and …
18e887d0
fix
88a7e8b4
typo
e9abe6ae
unify io binding api with non io binding
c9b45ee3
force evaluated shape to int
aad9aafe
mark pix2struct io binding tests
a29706ec
force contiguity in forward pass
821c997c
fixed cryptic contiguity problems
cc2e1248
fix some
3a2bcee4
fix vision2seq modeling and testing
f0ea2889
IlyasMoutawwakil
commented
on 2025-01-22
echarlaix
commented
on 2025-01-27
Merge branch 'main' into enable-ort-gpu-tests
7e122c0e
IlyasMoutawwakil
commented
on 2025-01-28
Update setup.py
dc2361db
update import utils
4eb95f10
Update optimum/onnxruntime/modeling_ort.py
7f1fc406
fix vision encoder decoder io binding
696cc95a
enable bigbird and bigbirg pegasus and seperate timm slow tests to un…
18274506
use bigger machine for slow tests
41abf7fe
lower atol and rtol for image classification logits
6f3084a5
fix
010030ea
large
445b2911
enable more Longformer and MCTCT
04c89040
enable commented models in export as well
18e1844e
uncomment timm slow models, big bird optimization and marian pkv comp…
4487c74c
Merge branch 'main' into enable-ort-gpu-tests
24d682e3
Merge branch 'main' into enable-ort-gpu-tests
def5fdb0
fix whisper/speech_to_text test and make convolution deterministic
458355d7
pin torch for ort training
881015c3
IlyasMoutawwakil
commented
on 2025-01-29
ctc and speech also uses convolution so has to be deterministic
7c8c56f5
revert vison2seq atol
3a4bac9b
Reviewers
Assignees
No one assigned
Labels
onnxruntime-gpu
onnxruntime-training
onnxruntime-slow
Milestone
No milestone
Login to write a write a comment.
Login via GitHub