fix(hub): pin pretrained artifacts to one commit

tests/policies/test_pretrained_revision_pinning.py

@@ -0,0 +1,122 @@
+# Copyright 2026 The HuggingFace Inc. team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# 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.
+
+import torch
+
+from lerobot.policies import factory
+from lerobot.policies.act.configuration_act import ACTConfig
+from lerobot.policies.pretrained import PreTrainedPolicy
+from lerobot.processor import PolicyProcessorPipeline
+
+
+class _MinimalPolicy(PreTrainedPolicy):
+    config_class = ACTConfig
+    name = "minimal_revision_test"
+
+    def get_optim_params(self) -> dict:
+        return {}
+
+    def reset(self) -> None:
+        pass
+
+    def forward(self, batch: dict[str, torch.Tensor]) -> tuple[torch.Tensor, dict | None]:
+        return torch.tensor(0), None
+
+    def predict_action_chunk(self, batch: dict[str, torch.Tensor], **kwargs) -> torch.Tensor:
+        return torch.tensor(0)
+
+    def select_action(self, batch: dict[str, torch.Tensor], **kwargs) -> torch.Tensor:
+        return torch.tensor(0)
+
+
+def _skip_safetensor_loading(monkeypatch):
+    monkeypatch.setattr(
+        _MinimalPolicy,
+        "_load_as_safetensor",
+        classmethod(lambda cls, model, model_file, map_location, strict: model),
+    )
+
+
+def test_policy_weights_use_config_commit_hash(monkeypatch):
+    config = ACTConfig(device="cpu")
+    config._set_hub_commit_hash("a" * 40, "user/policy")
+    calls = []
+
+    def fake_hub_download(**kwargs):
+        calls.append(kwargs)
+        return "/unused/model.safetensors"
+
+    monkeypatch.setattr("lerobot.policies.pretrained.hf_hub_download", fake_hub_download)
+    _skip_safetensor_loading(monkeypatch)
+
+    _MinimalPolicy.from_pretrained("user/policy", config=config, revision="main")
+
+    assert calls[0]["revision"] == "a" * 40
+
+
+def test_policy_does_not_reuse_commit_hash_for_another_repo(monkeypatch):
+    config = ACTConfig(device="cpu")
+    config._set_hub_commit_hash("a" * 40, "user/adapter")
+    calls = []
+
+    def fake_hub_download(**kwargs):
+        calls.append(kwargs)
+        return "/unused/model.safetensors"
+
+    monkeypatch.setattr("lerobot.policies.pretrained.hf_hub_download", fake_hub_download)
+    _skip_safetensor_loading(monkeypatch)
+
+    _MinimalPolicy.from_pretrained("user/base-policy", config=config, revision="base-tag")
+
+    assert calls[0]["revision"] == "base-tag"
+
+
+def test_policy_records_weight_commit_for_explicit_config(monkeypatch):
+    commit_hash = "a" * 40
+    config = ACTConfig(device="cpu")
+
+    def fake_hub_download(**kwargs):
+        return f"/cache/models--user--policy/snapshots/{commit_hash}/model.safetensors"
+
+    monkeypatch.setattr("lerobot.policies.pretrained.hf_hub_download", fake_hub_download)
+    _skip_safetensor_loading(monkeypatch)
+
+    _MinimalPolicy.from_pretrained("user/policy", config=config, revision="main")
+
+    assert config._commit_hash == commit_hash
+    assert config._commit_hash_source == "user/policy"
+
+
+def test_processor_factory_uses_config_commit_hash(monkeypatch):
+    config = ACTConfig(device="cpu")
+    config._set_hub_commit_hash("a" * 40, "user/policy")
+    calls = []
+
+    def fake_from_pretrained(cls, **kwargs):
+        calls.append(kwargs)
+        return PolicyProcessorPipeline(steps=[])
+
+    monkeypatch.setattr(
+        factory.PolicyProcessorPipeline,
+        "from_pretrained",
+        classmethod(fake_from_pretrained),
+    )
+
+    factory.make_pre_post_processors(
+        config,
+        pretrained_path="user/policy",
+        pretrained_revision="main",
+    )
+
+    assert [call["revision"] for call in calls] == ["a" * 40, "a" * 40]