Update

[ghstack-poisoned]

Author: Vincent Moens

benchmarks/test_objectives_benchmarks.py

@@ -6,6 +6,7 @@
 
 import pytest
 import torch
+from packaging import version
 
 from tensordict import TensorDict
 from tensordict.nn import (
@@ -43,6 +44,11 @@
     vec_td_lambda_return_estimate,
 )
 
+TORCH_VERSION = torch.__version__
+FULLGRAPH = version.parse(".".join(TORCH_VERSION.split(".")[:3])) >= version.parse(
+    "2.5.0"
+)  # Anything from 2.5, incl. nightlies, allows for fullgraph
+
 
 @pytest.fixture(scope="module")
 def set_default_device():
@@ -147,7 +153,7 @@ def test_gae_speed(benchmark, gae_fn, gamma_tensor, batches, timesteps):
     )
 
 
-def _maybe_compile(fn, compile, td, fullgraph=True, warmup=3):
+def _maybe_compile(fn, compile, td, fullgraph=FULLGRAPH, warmup=3):
     if compile:
         if isinstance(compile, str):
             fn = torch.compile(fn, mode=compile, fullgraph=fullgraph)

test/test_rb.py

@@ -546,6 +546,20 @@ def test_errors(self, storage_type):
         ):
             storage_type(data, max_size=4)
 
+    def test_existsok_lazymemmap(self, tmpdir):
+        storage0 = LazyMemmapStorage(10, scratch_dir=tmpdir)
+        rb = ReplayBuffer(storage=storage0)
+        rb.extend(TensorDict(a=torch.randn(3), batch_size=[3]))
+
+        storage1 = LazyMemmapStorage(10, scratch_dir=tmpdir)
+        rb = ReplayBuffer(storage=storage1)
+        with pytest.raises(RuntimeError, match="existsok"):
+            rb.extend(TensorDict(a=torch.randn(3), batch_size=[3]))
+
+        storage2 = LazyMemmapStorage(10, scratch_dir=tmpdir, existsok=True)
+        rb = ReplayBuffer(storage=storage2)
+        rb.extend(TensorDict(a=torch.randn(3), batch_size=[3]))
+
     @pytest.mark.parametrize(
         "data_type", ["tensor", "tensordict", "tensorclass", "pytree"]
     )

torchrl/data/replay_buffers/storages.py

@@ -923,6 +923,8 @@ class LazyMemmapStorage(LazyTensorStorage):
     Args:
         max_size (int): size of the storage, i.e. maximum number of elements stored
             in the buffer.
+
+    Keyword Args:
         scratch_dir (str or path): directory where memmap-tensors will be written.
         device (torch.device, optional): device where the sampled tensors will be
             stored and sent. Default is :obj:`torch.device("cpu")`.
@@ -933,6 +935,9 @@ class LazyMemmapStorage(LazyTensorStorage):
             measuring the storage size. For instance, a storage of shape ``[3, 4]``
             has capacity ``3`` if ``ndim=1`` and ``12`` if ``ndim=2``.
             Defaults to ``1``.
+        existsok (bool, optional): whether an error should be raised if any of the
+            tensors already exists on disk. Defaults to ``True``. If ``False``, the
+            tensor will be opened as is, not overewritten.
 
     .. note:: When checkpointing a ``LazyMemmapStorage``, one can provide a path identical to where the storage is
         already stored to avoid executing long copies of data that is already stored on disk.
@@ -1009,10 +1014,12 @@ def __init__(
         scratch_dir=None,
         device: torch.device = "cpu",
         ndim: int = 1,
+        existsok: bool = False,
     ):
         super().__init__(max_size, ndim=ndim)
         self.initialized = False
         self.scratch_dir = None
+        self.existsok = existsok
         if scratch_dir is not None:
             self.scratch_dir = str(scratch_dir)
             if self.scratch_dir[-1] != "/":
@@ -1108,7 +1115,7 @@ def max_size_along_dim0(data_shape):
         if is_tensor_collection(data):
             out = data.clone().to(self.device)
             out = out.expand(max_size_along_dim0(data.shape))
-            out = out.memmap_like(prefix=self.scratch_dir)
+            out = out.memmap_like(prefix=self.scratch_dir, existsok=self.existsok)
             for key, tensor in sorted(
                 out.items(include_nested=True, leaves_only=True), key=str
             ):