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

Jump to bottom

[Feature] Warn when reset_parameters_recursive is a no-op #693

Merged
merged 7 commits into from
Feb 26, 2024
Merged

[Feature] Warn when reset_parameters_recursive is a no-op #693

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
ddffb2e
amend
matteobettini Feb 26, 2024
35c5889
amend
matteobettini Feb 26, 2024
ba4412b
amend
matteobettini Feb 26, 2024
ba07d95
test
matteobettini Feb 26, 2024
169181c
Update tensordict/nn/common.py
vmoens Feb 26, 2024
d893c37
Update tensordict/nn/common.py
vmoens Feb 26, 2024
46f12a4
lint
vmoens Feb 26, 2024
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
17 changes: 11 additions & 6 deletions tensordict/nn/common.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,18 +13,17 @@

import torch
from cloudpickle import dumps as cloudpickle_dumps, loads as cloudpickle_loads

from tensordict._td import is_tensor_collection, TensorDictBase
from tensordict._tensordict import _unravel_key_to_tuple, unravel_key_list
from tensordict.functional import make_tensordict

from tensordict.nn.functional_modules import (
_swap_state,
extract_weights_and_buffers,
is_functional,
make_functional,
repopulate_module,
)

from tensordict.nn.utils import (
_auto_make_functional,
_dispatch_td_nn_modules,
Expand Down Expand Up @@ -248,7 +247,6 @@ def __call__(self, func: Callable) -> Callable:

@functools.wraps(func)
def wrapper(_self, *args: Any, **kwargs: Any) -> Any:

if not _dispatch_td_nn_modules():
return func(_self, *args, **kwargs)

Expand Down Expand Up @@ -830,7 +828,11 @@ def reset_parameters_recursive(
False
"""
if parameters is None:
self._reset_parameters(self)
any_reset = self._reset_parameters(self)
if not any_reset:
warnings.warn(
"reset_parameters_recursive was called without the parameters argument and did not find any parameters to reset"
)
return
elif parameters.ndim:
raise RuntimeError(
Expand Down Expand Up @@ -868,13 +870,16 @@ def reset_parameters_recursive(
self._reset_parameters(self)
return sanitized_parameters

def _reset_parameters(self, module: nn.Module) -> None:
def _reset_parameters(self, module: nn.Module) -> bool:
any_reset = False
for child in module.children():
if isinstance(child, nn.Module):
self._reset_parameters(child)
any_reset |= self._reset_parameters(child)

if hasattr(child, "reset_parameters"):
child.reset_parameters()
any_reset |= True
return any_reset


class TensorDictModule(TensorDictModuleBase):
Expand Down
56 changes: 42 additions & 14 deletions test/test_nn.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -42,7 +42,7 @@
set_skip_existing,
skip_existing,
)
from torch import distributions as d, nn
from torch import distributions, nn
from torch.distributions import Normal
from torch.utils._pytree import tree_map

Expand Down Expand Up @@ -165,6 +165,16 @@ def test_reset(self):
seq.reset_parameters_recursive()
assert torch.all(old_param != net[0][0].weight.data)

def test_reset_warning(self):
torch.manual_seed(0)
net = nn.ModuleList([nn.Tanh(), nn.ReLU()])
module = TensorDictModule(net, in_keys=["in"], out_keys=["out"])
with pytest.warns(
UserWarning,
match="reset_parameters_recursive was called without the parameters argument and did not find any parameters to reset",
):
module.reset_parameters_recursive()

@pytest.mark.parametrize(
"net",
[
Expand Down Expand Up @@ -335,7 +345,7 @@ def test_stateful_probabilistic_kwargs(
net = TensorDictModule(module=net, in_keys=in_keys, out_keys=out_keys)

kwargs = {
"distribution_class": torch.distributions.Uniform,
"distribution_class": distributions.Uniform,
"distribution_kwargs": {"high": max_dist},
}
if out_keys == ["low"]:
Expand Down Expand Up @@ -2666,7 +2676,6 @@ def test_module_buffer():
],
)
def test_nested_keys_probabilistic_delta(log_prob_key):

policy_module = TensorDictModule(
nn.Linear(1, 1), in_keys=[("data", "states")], out_keys=[("data", "param")]
)
Expand Down Expand Up @@ -2711,7 +2720,6 @@ def test_nested_keys_probabilistic_delta(log_prob_key):
],
)
def test_nested_keys_probabilistic_normal(log_prob_key):

loc_module = TensorDictModule(
nn.Linear(1, 1),
in_keys=[("data", "states")],
Expand Down Expand Up @@ -3083,7 +3091,10 @@ def test_const(self):
)
dist = CompositeDistribution(
params,
distribution_map={"cont": d.Normal, ("nested", "disc"): d.Categorical},
distribution_map={
"cont": distributions.Normal,
("nested", "disc"): distributions.Categorical,
},
)
assert dist.batch_shape == params.shape
assert len(dist.dists) == 2
Expand All @@ -3100,7 +3111,10 @@ def test_sample(self):
)
dist = CompositeDistribution(
params,
distribution_map={"cont": d.Normal, ("nested", "disc"): d.Categorical},
distribution_map={
"cont": distributions.Normal,
("nested", "disc"): distributions.Categorical,
},
)
sample = dist.sample()
assert sample.shape == params.shape
Expand All @@ -3121,8 +3135,8 @@ def test_rsample(self):
dist = CompositeDistribution(
params,
distribution_map={
"cont": d.Normal,
("nested", "disc"): d.RelaxedOneHotCategorical,
"cont": distributions.Normal,
("nested", "disc"): distributions.RelaxedOneHotCategorical,
},
extra_kwargs={("nested", "disc"): {"temperature": torch.tensor(1.0)}},
)
Expand All @@ -3147,8 +3161,8 @@ def test_log_prob(self):
dist = CompositeDistribution(
params,
distribution_map={
"cont": d.Normal,
("nested", "disc"): d.RelaxedOneHotCategorical,
"cont": distributions.Normal,
("nested", "disc"): distributions.RelaxedOneHotCategorical,
},
extra_kwargs={("nested", "disc"): {"temperature": torch.tensor(1.0)}},
)
Expand All @@ -3172,7 +3186,11 @@ def test_cdf(self):
[3],
)
dist = CompositeDistribution(
params, distribution_map={"cont": d.Normal, ("nested", "cont"): d.Normal}
params,
distribution_map={
"cont": distributions.Normal,
("nested", "cont"): distributions.Normal,
},
)
sample = dist.rsample((4,))
sample = dist.cdf(sample)
Expand All @@ -3194,7 +3212,11 @@ def test_icdf(self):
[3],
)
dist = CompositeDistribution(
params, distribution_map={"cont": d.Normal, ("nested", "cont"): d.Normal}
params,
distribution_map={
"cont": distributions.Normal,
("nested", "cont"): distributions.Normal,
},
)
sample = dist.rsample((4,))
sample = dist.cdf(sample)
Expand Down Expand Up @@ -3225,7 +3247,10 @@ def test_prob_module(self, interaction, return_log_prob):
)
in_keys = ["params"]
out_keys = ["cont", ("nested", "cont")]
distribution_map = {"cont": d.Normal, ("nested", "cont"): d.Normal}
distribution_map = {
"cont": distributions.Normal,
("nested", "cont"): distributions.Normal,
}
module = ProbabilisticTensorDictModule(
in_keys=in_keys,
out_keys=out_keys,
Expand Down Expand Up @@ -3275,7 +3300,10 @@ def test_prob_module_seq(self, interaction, return_log_prob):
)
in_keys = ["params"]
out_keys = ["cont", ("nested", "cont")]
distribution_map = {"cont": d.Normal, ("nested", "cont"): d.Normal}
distribution_map = {
"cont": distributions.Normal,
("nested", "cont"): distributions.Normal,
}
backbone = TensorDictModule(lambda: None, in_keys=[], out_keys=[])
module = ProbabilisticTensorDictSequential(
backbone,
Expand Down
Loading