Skip to content

[Unity][Frontend][NN] Switch to always using fp32 for timestep embedding calculation #15811

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

Merged
junrushao merged 2 commits into from
Sep 23, 2023
Merged

[Unity][Frontend][NN] Switch to always using fp32 for timestep embedding calculation #15811

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
5117538
Switch to always using fp32 for timestep embedding calculation.
Sep 22, 2023
5902da6
bump ci
Sep 23, 2023
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
15 changes: 10 additions & 5 deletions python/tvm/relax/frontend/nn/op.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1220,19 +1220,21 @@ def get_timestep_embedding(
[N x dim] Tensor of positional embeddings.
"""
dtype = get_default_dtype()
timesteps = _op.astype(x._expr, dtype)

# Arithmetic should be done in float for precision.
timesteps = _op.astype(x._expr, "float32")

half_dim = embedding_dim // 2
exponent = rx.const(-math.log(max_period), dtype) * _op.arange(
start=0, end=half_dim, dtype=dtype
exponent = rx.const(-math.log(max_period), "float32") * _op.arange(
start=0, end=half_dim, dtype="float32"
)
exponent = exponent / (rx.const(half_dim - downscale_freq_shift, dtype))
exponent = exponent / (rx.const(half_dim - downscale_freq_shift, "float32"))

emb = _op.exp(exponent)
emb = _op.expand_dims(timesteps, 1) * _op.expand_dims(emb, 0)
# Scale embeddings
if scale != 1:
emb = rx.const(scale, dtype) * emb
emb = rx.const(scale, "float32") * emb

# Concat sine and cosine embeddings.
if flip_sin_to_cos:
Expand All @@ -1243,6 +1245,9 @@ def get_timestep_embedding(
# Zero pad
if embedding_dim % 2 == 1:
emb = _op.nn.pad(emb, (0, 1, 0, 0))

# Cast to proper output type
emb = _op.astype(emb, dtype)
return _wrap_nested(emb, name)


Expand Down
5 changes: 3 additions & 2 deletions tests/python/relax/test_frontend_nn_modules.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -402,8 +402,9 @@ def forward(
lv8: R.Tensor((3, 5), dtype="float32") = R.multiply(lv2, lv7)
lv9: R.Tensor((3, 5), dtype="float32") = R.sin(lv8)
lv10: R.Tensor((3, 5), dtype="float32") = R.cos(lv8)
get_timestep_embedding: R.Tensor((3, 10), dtype="float32") = R.concat(
(lv9, lv10), axis=-1
lv11: R.Tensor((3, 10), dtype="float32") = R.concat((lv9, lv10), axis=-1)
get_timestep_embedding: R.Tensor((3, 10), dtype="float32") = R.astype(
lv11, dtype="float32"
)
gv1: R.Tuple(
R.Tensor((3, 10), dtype="float32"), R.Tuple(R.Object)
Expand Down
5 changes: 3 additions & 2 deletions tests/python/relax/test_frontend_nn_op.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -386,8 +386,9 @@ def test(
lv8: R.Tensor((3, 5), dtype="float32") = R.multiply(lv2, lv7)
lv9: R.Tensor((3, 5), dtype="float32") = R.sin(lv8)
lv10: R.Tensor((3, 5), dtype="float32") = R.cos(lv8)
get_timestep_embedding: R.Tensor((3, 10), dtype="float32") = R.concat(
(lv9, lv10), axis=-1
lv11: R.Tensor((3, 10), dtype="float32") = R.concat((lv9, lv10), axis=-1)
get_timestep_embedding: R.Tensor((3, 10), dtype="float32") = R.astype(
lv11, dtype="float32"
)
gv1: R.Tuple(
R.Tensor((3, 10), dtype="float32"), R.Tuple(R.Object)
Expand Down