Merge remote-tracking branch 'ParamThakkar123/add/procgen' into add/procgen

Author: vmoens

.github/workflows/test-linux-libs.yml

@@ -541,7 +541,7 @@ jobs:
   unittests-procgen:
     strategy:
       matrix:
-        python_version: ["3.9"]
+        python_version: ["3.10"]
         cuda_arch_version: ["12.8"]
     if: ${{ github.event_name == 'push' || contains(github.event.pull_request.labels.*.name, 'Environments') }}
     uses: pytorch/test-infra/.github/workflows/linux_job_v2.yml@main

test/test_libs.py

@@ -5416,61 +5416,6 @@ def test_isaaclab_reset(self, env):
         assert not r["next", "policy"][r["next", "done"].squeeze(-1)].isfinite().any()
 
 
-@pytest.mark.skipif(not _has_procgen, reason="Procgen not found")
-class TestProcgen:
-    @pytest.mark.parametrize("envname", ["coinrun", "starpilot"])
-    def test_procgen_envs_available(self, envname):
-        # availability check
-        assert envname in ProcgenEnv.available_envs
-
-    def test_procgen_invalid_env_raises(self):
-        with pytest.raises(ValueError):
-            ProcgenEnv("this_env_does_not_exist")
-
-    def test_procgen_num_envs_batch_size(self):
-        env = ProcgenEnv("coinrun", num_envs=3)
-        td = env.reset()
-        assert td["observation"].shape[0] == 3
-        env.close()
-
-    def test_procgen_seeding_is_deterministic(self):
-        e1 = ProcgenEnv("coinrun", num_envs=2)
-        e2 = ProcgenEnv("coinrun", num_envs=2)
-        e1.set_seed(0)
-        e2.set_seed(0)
-        t1 = e1.reset()
-        t2 = e2.reset()
-        assert torch.equal(t1["observation"], t2["observation"])
-        e1.close()
-        e2.close()
-
-    def test_procgen_step_keys_and_shapes(self):
-        env = ProcgenEnv("coinrun", num_envs=2)
-        env.reset()
-        td = env.rand_step()
-        for k in ("observation", "reward", "done"):
-            assert k in td
-        assert td["observation"].shape[0] == 2
-        env.close()
-
-    @pytest.mark.skipif(not _has_procgen, reason="Procgen not found")
-    def test_procgen_env_creation_and_reset(self):
-        env = ProcgenEnv("coinrun", num_envs=4)
-        td = env.reset()
-        # ensure batch size corresponds to num_envs
-        assert td["observation"].shape[0] == 4
-
-    @pytest.mark.skipif(not _has_procgen, reason="Procgen not found")
-    def test_procgen_env_step(self):
-        env = ProcgenEnv("coinrun", num_envs=2)
-        env.reset()
-        out = env.rand_step()
-        # basic checks on returned tensordict
-        assert "observation" in out
-        assert "reward" in out
-        assert "done" in out
-
-
 @pytest.mark.skipif(not _has_procgen, reason="Procgen not found")
 class TestProcgen:
     @pytest.mark.parametrize("envname", ["coinrun", "starpilot"])