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.

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

[nnx] add Dropout.rngs #3815

Merged
merged 1 commit into from
Apr 9, 2024
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
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
10 changes: 0 additions & 10 deletions flax/experimental/nnx/nnx/module.py
Original file line number Diff line number Diff line change
Expand Up @@ -68,16 +68,6 @@ def _module_meta_call(cls: tp.Type[M], *args, **kwargs) -> M:
if isinstance(module, _HasSetup):
module.setup()

assert isinstance(module, Module)

for field in dataclasses.fields(module):
if not field.init:
continue
value = vars(module)[field.name]
# set Rngs instances to None
if isinstance(value, Rngs):
vars(module)[field.name] = None

return module


Expand Down
13 changes: 8 additions & 5 deletions flax/experimental/nnx/nnx/nn/stochastic.py
Original file line number Diff line number Diff line change
Expand Up @@ -11,16 +11,17 @@
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from __future__ import annotations

from typing import Optional, Sequence
import dataclasses
from typing import Sequence

import jax
import jax.numpy as jnp
from jax import lax, random

from flax.experimental.nnx.nnx import rnglib
from flax.experimental.nnx.nnx.module import Module, first_from
import dataclasses


@dataclasses.dataclass
Expand All @@ -38,15 +39,16 @@ class Dropout(Module):

rate: float
broadcast_dims: Sequence[int] = ()
deterministic: Optional[bool] = None
deterministic: bool | None = None
rng_collection: str = 'dropout'
rngs: rnglib.Rngs | None = None

def __call__(
self,
inputs,
*,
deterministic: Optional[bool] = None,
rngs: Optional[rnglib.Rngs] = None,
deterministic: bool | None = None,
rngs: rnglib.Rngs | None = None,
) -> jax.Array:
"""Applies a random dropout mask to the input.

Expand All @@ -59,6 +61,7 @@ def __call__(
Returns:
The masked inputs reweighted to preserve mean.
"""
rngs = rngs or self.rngs
deterministic = first_from(
deterministic,
self.deterministic,
Expand Down
27 changes: 27 additions & 0 deletions flax/experimental/nnx/tests/nn/test_stochastic.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,27 @@

import jax.numpy as jnp

from flax.experimental import nnx


class TestStochastic:
def test_dropout_internal_rngs(self):
n = 0
m = nnx.Dropout(rate=0.5, deterministic=False, rngs=nnx.Rngs(dropout=0))

@nnx.jit
def f(m, x):
nonlocal n
n += 1
return m(x)

x = jnp.ones((1, 10))
assert m.rngs is not None and m.rngs.dropout.count.value == 0

y = f(m, x)
assert n == 1
assert m.rngs.dropout.count.value == 1

y = f(m, x)
assert n == 1
assert m.rngs.dropout.count.value == 2
4 changes: 1 addition & 3 deletions flax/experimental/nnx/tests/test_module.py
Original file line number Diff line number Diff line change
Expand Up @@ -550,7 +550,7 @@ class Foo(nnx.Module):
assert state.d == nnx.Variable(4)
assert state.e == nnx.BatchStat(5)

def test_context_none_after_init(self):
def test_post_init(self):
@dataclasses.dataclass
class DFoo(nnx.Module):
din: int
Expand All @@ -566,7 +566,6 @@ def __call__(self, x):
m = DFoo(1, 1, rngs=nnx.Rngs(0))

assert hasattr(m, 'bar')
assert m.rngs is None

def test_setup_is_called(self):
@dataclasses.dataclass
Expand All @@ -584,7 +583,6 @@ def __call__(self, x):
m = DFoo(1, 1, rngs=nnx.Rngs(0))

assert hasattr(m, 'bar')
assert m.rngs is None


class TestModuleDef:
Expand Down