grid search test

examples/grid_search_usage.py

@@ -28,6 +28,7 @@
 Notes:
     1. local_mode must set to (False) to enable parallelized training;
     2. other tunable parameters can be found in model arch and algorithm config
+    3. grid_search function not available on: trpo family and joint Q family
 """
 
 from marllib import marl

tests/test_grid_search.py

@@ -0,0 +1,163 @@
+# MIT License
+
+# Copyright (c) 2023 Replicable-MARL
+
+# Permission is hereby granted, free of charge, to any person obtaining a copy
+# of this software and associated documentation files (the "Software"), to deal
+# in the Software without restriction, including without limitation the rights
+# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+# copies of the Software, and to permit persons to whom the Software is
+# furnished to do so, subject to the following conditions:
+#
+# The above copyright notice and this permission notice shall be included in all
+# copies or substantial portions of the Software.
+#
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+# SOFTWARE.
+
+import unittest
+from marllib import marl
+from ray import tune
+
+
+class TestBaseOnMPE(unittest.TestCase):
+
+    #################
+    ### algorithm ###
+    #################
+
+    # HA algorithm
+    def test_a1_hatrpo(self):  # search not success
+        env = marl.make_env(environment_name="mpe", map_name="simple_spread", force_coop=True)
+        algo = marl.algos.hatrpo(hyperparam_source="test", critic_lr=tune.grid_search([0.0005, 0.001]))
+        model = marl.build_model(env, algo, {"core_arch": "mlp", "encode_layer": "8-8"})
+        algo.fit(env, model, stop={"training_iteration": 1}, local_mode=False, num_gpus=0,
+                 num_workers=2, share_policy="individual", checkpoint_end=False)
+
+    def test_a2_happo(self):
+        env = marl.make_env(environment_name="mpe", map_name="simple_spread", force_coop=True)
+        algo = marl.algos.happo(hyperparam_source="test", lr=tune.grid_search([0.0005, 0.001]))
+        model = marl.build_model(env, algo, {"core_arch": "gru", "encode_layer": "8-8"})
+        algo.fit(env, model, stop={"training_iteration": 1}, local_mode=False, num_gpus=0,
+                 num_workers=2, share_policy="individual", checkpoint_end=False)
+
+    def test_b1_maa2c(self):
+        env = marl.make_env(environment_name="mpe", map_name="simple_adversary")
+        algo = marl.algos.maa2c(hyperparam_source="test", lr=tune.grid_search([0.0005, 0.001]))
+        model = marl.build_model(env, algo, {"core_arch": "lstm", "encode_layer": "8-8"})
+        algo.fit(env, model, stop={"training_iteration": 1}, local_mode=False, num_gpus=0,
+                 num_workers=2, share_policy="group", checkpoint_end=False)
+
+    def test_b2_coma(self):
+        env = marl.make_env(environment_name="mpe", map_name="simple_adversary", continuous_actions=False)
+        algo = marl.algos.coma(hyperparam_source="test", lr=tune.grid_search([0.0005, 0.001]))
+        model = marl.build_model(env, algo, {"core_arch": "lstm", "encode_layer": "8-8"})
+        algo.fit(env, model, stop={"training_iteration": 1}, local_mode=False, num_gpus=0,
+                 num_workers=2, share_policy="group", checkpoint_end=False)
+
+    def test_b3_matrpo(self):   # search not success
+        env = marl.make_env(environment_name="mpe", map_name="simple_adversary")
+        algo = marl.algos.matrpo(hyperparam_source="test", critic_lr=tune.grid_search([0.0005, 0.001]))
+        model = marl.build_model(env, algo, {"core_arch": "lstm", "encode_layer": "8-8"})
+        algo.fit(env, model, stop={"training_iteration": 1}, local_mode=False, num_gpus=0,
+                 num_workers=2, share_policy="group", checkpoint_end=False)
+
+    def test_b4_mappo(self):
+        env = marl.make_env(environment_name="mpe", map_name="simple_spread")
+        algo = marl.algos.mappo(hyperparam_source="test", lr=tune.grid_search([0.0005, 0.001]))
+        model = marl.build_model(env, algo, {"core_arch": "gru", "encode_layer": "8-8"})
+        algo.fit(env, model, stop={"training_iteration": 1}, local_mode=False, num_gpus=0,
+                 num_workers=2, share_policy="individual", checkpoint_end=False)
+
+    def test_b51_maddpg(self):
+        env = marl.make_env(environment_name="mpe", map_name="simple_spread", continuous_actions=True)
+        algo = marl.algos.maddpg(hyperparam_source="test", critic_lr=tune.grid_search([0.0005, 0.001]))
+        model = marl.build_model(env, algo, {"core_arch": "mlp", "encode_layer": "8-8"})
+        algo.fit(env, model, stop={"training_iteration": 1}, local_mode=False, num_gpus=0,
+                 num_workers=2, share_policy="all", checkpoint_end=False)
+
+    # IL algorithm
+    def test_c1_ia2c(self):
+        env = marl.make_env(environment_name="mpe", map_name="simple_spread")
+        algo = marl.algos.ia2c(hyperparam_source="test", lr=tune.grid_search([0.0005, 0.001]))
+        model = marl.build_model(env, algo, {"core_arch": "gru", "encode_layer": "8-8"})
+        algo.fit(env, model, stop={"training_iteration": 1}, local_mode=False, num_gpus=0,
+                 num_workers=2, share_policy="all", checkpoint_end=False)
+
+    def test_c2_itrpo(self): # search not success
+        env = marl.make_env(environment_name="mpe", map_name="simple_spread")
+        algo = marl.algos.itrpo(hyperparam_source="test", critic_lr=tune.grid_search([0.0005, 0.001]))
+        model = marl.build_model(env, algo, {"core_arch": "lstm", "encode_layer": "8-8"})
+        algo.fit(env, model, stop={"training_iteration": 1}, local_mode=False, num_gpus=0,
+                 num_workers=2, share_policy="group", checkpoint_end=False)
+
+    def test_c3_ippo(self):
+        env = marl.make_env(environment_name="mpe", map_name="simple_spread")
+        algo = marl.algos.ippo(hyperparam_source="test", lr=tune.grid_search([0.0005, 0.001]))
+        model = marl.build_model(env, algo, {"core_arch": "lstm", "encode_layer": "8-8"})
+        algo.fit(env, model, stop={"training_iteration": 1}, local_mode=False, num_gpus=0,
+                 num_workers=2, share_policy="individual", checkpoint_end=False)
+
+    def test_c3_iddpg(self):
+        env = marl.make_env(environment_name="mpe", map_name="simple_spread", continuous_actions=True)
+        algo = marl.algos.iddpg(hyperparam_source="test", critic_lr=tune.grid_search([0.0005, 0.001]))
+        model = marl.build_model(env, algo, {"core_arch": "gru", "encode_layer": "8-8"})
+        algo.fit(env, model, stop={"training_iteration": 1}, local_mode=False, num_gpus=0,
+                 num_workers=2, share_policy="all", checkpoint_end=False)
+
+
+    def test_c4_iql(self): # search not success
+        env = marl.make_env(environment_name="mpe", map_name="simple_spread", force_coop=True, continuous_actions=False)
+        algo = marl.algos.iql(hyperparam_source="test", lr=tune.grid_search([0.0005, 0.001]))
+        model = marl.build_model(env, algo, {"core_arch": "gru", "encode_layer": "8-8"})
+        algo.fit(env, model, stop={"training_iteration": 1}, local_mode=False, num_gpus=0,
+                 num_workers=2, share_policy="all", checkpoint_end=False)
+
+    # VD algorithms
+
+    def test_d1_qmix(self): # search not success
+        env = marl.make_env(environment_name="mpe", map_name="simple_spread", force_coop=True, continuous_actions=False)
+        algo = marl.algos.qmix(hyperparam_source="test", lr=tune.grid_search([0.0005, 0.001]))
+        model = marl.build_model(env, algo, {"core_arch": "gru", "encode_layer": "8-8"})
+        algo.fit(env, model, stop={"training_iteration": 1}, local_mode=False, num_gpus=0,
+                 num_workers=2, share_policy="all", checkpoint_end=False)
+
+    def test_d2_vdn(self): # search not success
+        env = marl.make_env(environment_name="mpe", map_name="simple_spread", force_coop=True, continuous_actions=False)
+        algo = marl.algos.vdn(hyperparam_source="test", lr=tune.grid_search([0.0005, 0.001]))
+        model = marl.build_model(env, algo, {"core_arch": "gru", "encode_layer": "8-8"})
+        algo.fit(env, model, stop={"training_iteration": 1}, local_mode=False, num_gpus=0,
+                 num_workers=2, share_policy="all", checkpoint_end=False)
+
+    def test_d3_vda2c(self):
+        env = marl.make_env(environment_name="mpe", map_name="simple_spread", force_coop=True, continuous_actions=True)
+        algo = marl.algos.vda2c(hyperparam_source="test", lr=tune.grid_search([0.0005, 0.001]))
+        model = marl.build_model(env, algo, {"core_arch": "gru", "encode_layer": "8-8"})
+        algo.fit(env, model, stop={"training_iteration": 1}, local_mode=False, num_gpus=0,
+                 num_workers=2, share_policy="group", checkpoint_end=False)
+
+    def test_d4_vdppo(self):
+        env = marl.make_env(environment_name="mpe", map_name="simple_spread", force_coop=True, continuous_actions=True)
+        algo = marl.algos.vdppo(hyperparam_source="test", lr=tune.grid_search([0.0005, 0.001]))
+        model = marl.build_model(env, algo, {"core_arch": "lstm", "encode_layer": "8-8"})
+        algo.fit(env, model, stop={"training_iteration": 1}, local_mode=False, num_gpus=0,
+                 num_workers=2, share_policy="individual", checkpoint_end=False)
+
+    def test_d5_facmac(self):
+        env = marl.make_env(environment_name="mpe", map_name="simple_spread", force_coop=True, continuous_actions=True)
+        algo = marl.algos.facmac(hyperparam_source="test", critic_lr=tune.grid_search([0.0005, 0.001]))
+        model = marl.build_model(env, algo, {"core_arch": "gru", "encode_layer": "8-8"})
+        algo.fit(env, model, stop={"training_iteration": 1}, local_mode=False, num_gpus=0,
+                 num_workers=2, share_policy="all", checkpoint_end=False)
+
+
+if __name__ == "__main__":
+    import pytest
+    import sys
+
+    sys.exit(pytest.main(["-v", __file__]))