Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Dynamic shape support for llama3 #7

Merged
merged 1 commit into from
Jun 12, 2024
Merged

Dynamic shape support for llama3 #7

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions src/bindings/python/src/openvino/frontend/pytorch/torchdynamo/op_support.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -241,6 +241,7 @@ def __init__(self, options):
"torch.ops.aten.transpose.int": None,
"torch.ops.aten.tril.default": None,
"torch.ops.aten.tril_.default": None,
"torch.ops.aten.triu.default": None,
"torch.ops.aten.unbind.int": None,
"torch.ops.aten.unfold.default": None,
"torch.ops.aten.unsqueeze.default": None,
Expand Down
4 changes: 4 additions & 0 deletions src/frontends/pytorch/src/op/arange.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,6 +7,7 @@
#include "openvino/op/convert.hpp"
#include "openvino/op/convert_like.hpp"
#include "openvino/op/range.hpp"
#include "openvino/op/squeeze.hpp"
#include "pt_framework_node.hpp"
#include "utils.hpp"

Expand Down Expand Up @@ -108,6 +109,9 @@ OutputVector translate_arange_fx(const NodeContext& context) {
if (context.has_attribute("dtype")) {
dtype = context.get_attribute<element::Type>("dtype");
}
if (end.get_partial_shape().rank().is_dynamic()) {
end = context.mark_node(std::make_shared<ov::op::v0::Squeeze>(end, zero));
}
auto range = context.mark_node(std::make_shared<v4::Range>(start, end, step, dtype));
if (!context.has_attribute("dtype")) {
range = context.mark_node(std::make_shared<v1::ConvertLike>(range, context.get_input(0)));
Expand Down
33 changes: 21 additions & 12 deletions src/frontends/pytorch/src/op/expand.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,6 +5,7 @@
#include "openvino/frontend/pytorch/node_context.hpp"
#include "openvino/op/abs.hpp"
#include "openvino/op/broadcast.hpp"
#include "openvino/op/concat.hpp"
#include "openvino/op/shape_of.hpp"
#include "utils.hpp"

Expand Down Expand Up @@ -46,23 +47,31 @@ OutputVector translate_expand_fx(const NodeContext& context) {
num_inputs_check(context, 2, num_inputs);
auto x = context.get_input(0);
std::vector<int32_t> shape_vec;
auto sizes = context.get_input(1);
if (num_inputs != 2) {
if (context.get_input_type(1).is<type::List>()) {
std::deque<Output<Node>> list_elems;
for (size_t i = 1; i < num_inputs; i++) {
auto a = context.get_input_from_visible_context(i).get_node_shared_ptr();
auto shape_input = context.get_input(static_cast<int>(i));
if (std::dynamic_pointer_cast<ov::op::v0::Parameter>(a) ||
shape_input.get_partial_shape().rank().is_dynamic() ||
shape_input.get_partial_shape().rank().get_length() == 0) {
shape_vec.push_back(-1);
if (context.get_input_type(i).as<type::List>().element_type.is<type::PyScalar>()) {
auto const_val = context.const_input<int32_t>(i);
std::vector<int32_t> dim_vec;
dim_vec.push_back(const_val);
auto dim_const = ov::op::v0::Constant::create(element::i32, Shape{1}, dim_vec);
list_elems.push_back(dim_const);
} else {
auto val = context.const_input<int32_t>(i);
shape_vec.push_back(val);
auto converted_dim = context.mark_node(std::make_shared<ov::op::v0::Convert>(context.get_input(static_cast<int>(i)), element::i32));
list_elems.push_back(converted_dim);
}
}
sizes = ov::op::v0::Constant::create(element::i32, Shape{num_inputs - 1}, shape_vec);
auto concat = std::make_shared<ov::op::v0::Concat>(OutputVector(list_elems.begin(), list_elems.end()), 0);
return base_expand(context, x, concat);
} else {
auto x = context.get_input(0);
auto sizes = context.get_input(1);
// TODO: figure out what implicit means
PYTORCH_OP_CONVERSION_CHECK(context.input_is_none(2) || context.const_input<bool>(2) == false,
"Unexpected value of implicit for expand operation");
return base_expand(context, x, sizes);
}
return base_expand(context, x, sizes);

};

} // namespace op
Expand Down
26 changes: 23 additions & 3 deletions src/frontends/pytorch/src/op/full.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,6 +8,7 @@
#include "openvino/op/constant.hpp"
#include "openvino/op/convert_like.hpp"
#include "openvino/op/divide.hpp"
#include "openvino/op/concat.hpp"
#include "openvino/op/gather.hpp"
#include "openvino/op/multiply.hpp"
#include "openvino/op/power.hpp"
Expand Down Expand Up @@ -74,9 +75,28 @@ OutputVector translate_full(const NodeContext& context) {
OutputVector translate_full_fx(const NodeContext& context) {
// aten.full.default([16, 16], 0, dtype = torch.float32, layout = torch.strided, device = device(type='cpu'),
// pin_memory = False)
num_inputs_check(context, 2, 2);
auto sizes = context.get_input(0);
auto value = context.get_input(1);
auto num_inputs = context.get_input_size();
num_inputs_check(context, 2, num_inputs);
ov::Output<ov::Node> sizes;
if (context.get_input_type(0).is<type::List>()) {
std::deque<Output<Node>> list_elems;
for (size_t i = 0; i < num_inputs-1; i++) {
if (context.get_input_type(i).as<type::List>().element_type.is<type::PyScalar>()) {
auto const_val = context.const_input<int32_t>(i);
std::vector<int32_t> dim_vec;
dim_vec.push_back(const_val);
auto dim_const = ov::op::v0::Constant::create(element::i32, Shape{1}, dim_vec);
list_elems.push_back(dim_const);
} else {
auto converted_dim = context.mark_node(std::make_shared<ov::op::v0::Convert>(context.get_input(static_cast<int>(i)), element::i32));
list_elems.push_back(converted_dim);
}
}
sizes = std::make_shared<ov::op::v0::Concat>(OutputVector(list_elems.begin(), list_elems.end()), 0);
} else {
sizes = context.get_input(0);
}
auto value = context.get_input(num_inputs-1);

auto filled_tensor = base_translate_full(context, sizes, value);
if (context.has_attribute("dtype")) {
Expand Down
52 changes: 9 additions & 43 deletions src/frontends/pytorch/src/op/reshape.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -31,55 +31,21 @@ OutputVector translate_reshape_fx(const NodeContext& context) {
num_inputs_check(context, 2, num_inputs);
std::vector<int32_t> shape_vec;
if (context.get_input_type(1).is<type::List>()) {
int num_dyn_dims = 0;
std::deque<Output<Node>> list_elems;
for (size_t i = 1; i < num_inputs; i++) {
auto shape_input = context.get_input(static_cast<int>(i));
if (context.get_input_type(i).as<type::List>().element_type.is<type::PyScalar>()) {
auto const_val = context.const_input<int32_t>(i);
shape_vec.push_back(const_val);
std::vector<int32_t> dim_vec;
dim_vec.push_back(const_val);
auto dim_const = ov::op::v0::Constant::create(element::i32, Shape{1}, dim_vec);
list_elems.push_back(dim_const);
} else {
// Set dimension to be dynamic if it's coming from an argument or another node
shape_vec.push_back(-1);
num_dyn_dims++;
auto converted_dim = context.mark_node(std::make_shared<ov::op::v0::Convert>(context.get_input(static_cast<int>(i)), element::i32));
list_elems.push_back(converted_dim);
}
}
// We cannot use multiple -1s if there are more than 1 dynamic dimensions
if (num_dyn_dims >= 2) {
auto inp_shape = context.get_input(0).get_partial_shape();
// If there are multiple dynamic dymensions, we cannot support inputs with dynamic rank
if (inp_shape.rank().is_static()) {
auto zero = context.mark_node(ov::op::v0::Constant::create(element::i32, Shape{1}, {0}));
if (inp_shape.size() >= 3 && inp_shape.size() + 1 == shape_vec.size() && shape_vec[0] == 1 &&
inp_shape[0] == shape_vec[1]) {
// [N, ...] -> [1, N, ...] Can be translated to Unsqueeze
auto unsqueeze =
context.mark_node(std::make_shared<ov::op::v0::Unsqueeze>(context.get_input(0), zero));
return {unsqueeze};
} else if (shape_vec.size() >= 3 && shape_vec.size() + 1 == inp_shape.size() && inp_shape[0] == 1 &&
inp_shape[1] == shape_vec[0]) {
// [1, N, ...] -> [N, ...] Can be translated to Squeeze
auto squeeze = context.mark_node(std::make_shared<ov::op::v0::Squeeze>(context.get_input(0), zero));
return {squeeze};
} else if (inp_shape.size() == shape_vec.size()) {
// If the input rank is equal to output rank, we can use 0s in place of dynamic dimensions
for (size_t k = 0; k < shape_vec.size(); k++) {
if (shape_vec[k] == -1)
shape_vec[k] = 0;
}
} else {
FRONT_END_GENERAL_CHECK(
false,
"Cannot support reshape with multiple dynamic dimensions for unequal ranks");
}
} else {
FRONT_END_GENERAL_CHECK(
false,
"Cannot support reshape with multiple dynamic dimensions for dynamic input ranks");
}
}

auto shape_const = ov::op::v0::Constant::create(element::i32, Shape{num_inputs - 1}, shape_vec);
auto reshape = std::make_shared<ov::op::v1::Reshape>(context.get_input(0), shape_const, true);
auto concat = std::make_shared<ov::op::v0::Concat>(OutputVector(list_elems.begin(), list_elems.end()), 0);
auto reshape = std::make_shared<ov::op::v1::Reshape>(context.get_input(0), concat, true);
return {context.mark_node(reshape)};
} else {
auto shape_input = context.get_input(1);
Expand Down
4 changes: 2 additions & 2 deletions src/frontends/pytorch/src/op/slice_scatter.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -44,7 +44,7 @@ OutputVector translate_slice_scatter_fx(const NodeContext& context) {
ov::Output<ov::Node> end;
if (!context.input_is_none(4)) {
end = context.get_input(4);
if (end.get_partial_shape().rank().is_dynamic() || end.get_partial_shape().rank().get_length() == 0) {
if (!(end.get_partial_shape().rank().is_dynamic()) && end.get_partial_shape().rank().get_length() == 0) {
end = context.mark_node(std::make_shared<v0::Unsqueeze>(end, axis_0));
}
} else {
Expand All @@ -65,4 +65,4 @@ OutputVector translate_slice_scatter_fx(const NodeContext& context) {
} // namespace op
} // namespace pytorch
} // namespace frontend
} // namespace ov
} // namespace ov
3 changes: 2 additions & 1 deletion src/frontends/pytorch/src/op_table.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -758,7 +758,7 @@ const std::map<std::string, CreatorFunction> get_supported_ops_fx() {
{"aten._scaled_dot_product_flash_attention_for_cpu.default", op::translate_scaled_dot_product_attention_fx},
{"aten._softmax.default", op::translate_softmax_fx},
{"aten._to_copy.default", op::translate_to_fx},
{"aten._unsafe_view.default", op::translate_reshape},
{"aten._unsafe_view.default", op::translate_reshape_fx},
{"aten.abs.default", op::translate_1to1_match_1_inputs<opset10::Abs>},
{"aten.acos.default", op::translate_1to1_match_1_inputs_with_fp32_type_alignment<opset10::Acos>},
{"aten.acosh.default", op::translate_1to1_match_1_inputs_with_fp32_type_alignment<opset10::Acosh>},
Expand Down Expand Up @@ -963,6 +963,7 @@ const std::map<std::string, CreatorFunction> get_supported_ops_fx() {
{"aten.topk.default", op::translate_topk_fx},
{"aten.transpose.int", op::translate_transpose},
{"aten.tril.default", op::translate_tril},
{"aten.triu.default", op::translate_triu},
{"aten.unbind.int", op::translate_unbind_int_fx},
{"aten.unfold.default", op::translate_unfold},
{"aten.unsqueeze.default", op::translate_1to1_match_2_inputs<opset10::Unsqueeze>},
Expand Down
Loading