Update codegen to use xla::Shape #4111
wonjoo-wj
force pushed
from
51335035
to
c8074d57
3 years ago
wonjoo-wj
changed the title [WIP] Update GenXlaLazyIR codegen to use xla::Shape Update GenXlaLazyIR codegen to use xla::Shape 3 years ago
wonjoo-wj
changed the title Update GenXlaLazyIR codegen to use xla::Shape Update codegen to use xla::Shape 3 years ago
wonjoo-wj
marked this pull request as ready for review 3 years ago
wonjoo-wj
force pushed
from
3d4fcc40
to
b7f64f2a
3 years ago
wonjoo-wj
force pushed
from
ff8714a0
to
a2cb03ee
3 years ago
Update GenXlaLazyIR codegen to use xla::Shape
daf91440
Add xla::Shape to custom shape class
5269e394
Revert "Update GenXlaLazyIR codegen to use xla::Shape"
e4d37476
Update xla codegen to have custom GenLazyNativeFuncDefinition generator
879e11e4
Override build_ir_node() in GenXlaLazyNativeFuncDefinition
f7ef7ca2
Override gen() in GenXlaLazyIr
af82f3c0
Update existing MakeNode calls to remove torch::lazy::Shape vector
054b99d0
Remove torch::lazy::Shape vector for _adaptive_avg_pool2d_* ops
d15b8b55
Update _adaptive_avg_pool3d_* ops as well
8f26fe6e
Run linter
e6848899
Add torch pin
0e00007d
Remove lazy::Shape related logic in XLATensor::Create
6802fb6a
Clean up scripts/gen_lazy_tensor.py and add some comments
7d6ef314
Update cpp tests to remove torch::lazy:Shape constructor calls
50497866
Run linter again
069adc0c
Move bitwise ops to ir_codegen
d22314b3
Use new flag use_lazy_shape in xla lazy codegen
ce10b905
Update torch pin
982692d5
wonjoo-wj
force pushed
from
f571565e
to
982692d5
3 years ago
Clean up test_symint.cpp
c79917c7
Run linter on python files
33d447d4
Delete .torch_pin
58986bcc
wonjoo-wj
merged
587a6b6d
into master 3 years ago
Labels
REMOVE_TORCH_PIN
tracing
Login to write a write a comment.
Login via GitHub