[RLlib] Fix deprecated warning for torch_ops.py (soft-replaced by torch_utils.py). (#19982)

rllib/agents/a3c/a3c_torch_policy.py

@@ -15,7 +15,7 @@ from ray.rllib.policy.torch_policy import LearningRateSchedule, \
     EntropyCoeffSchedule
 from ray.rllib.utils.deprecation import Deprecated
 from ray.rllib.utils.framework import try_import_torch
-from ray.rllib.utils.torch_ops import apply_grad_clipping, sequence_mask
+from ray.rllib.utils.torch_utils import apply_grad_clipping, sequence_mask
 from ray.rllib.utils.typing import TrainerConfigDict, TensorType, \
     PolicyID, LocalOptimizer
 

rllib/agents/ars/ars.py

@@ -18,7 +18,7 @@ from ray.rllib.env.env_context import EnvContext
 from ray.rllib.policy.sample_batch import DEFAULT_POLICY_ID
 from ray.rllib.utils.annotations import override
 from ray.rllib.utils.deprecation import Deprecated
-from ray.rllib.utils.torch_ops import set_torch_seed
+from ray.rllib.utils.torch_utils import set_torch_seed
 from ray.rllib.utils import FilterManager
 
 logger = logging.getLogger(__name__)

rllib/agents/cql/cql_torch_policy.py

@@ -22,7 +22,7 @@ from ray.rllib.utils.framework import try_import_torch
 from ray.rllib.utils.metrics.learner_info import LEARNER_STATS_KEY
 from ray.rllib.utils.typing import LocalOptimizer, TensorType, \
     TrainerConfigDict
-from ray.rllib.utils.torch_ops import apply_grad_clipping, \
+from ray.rllib.utils.torch_utils import apply_grad_clipping, \
     convert_to_torch_tensor, concat_multi_gpu_td_errors
 
 torch, nn = try_import_torch()

rllib/agents/ddpg/ddpg_torch_policy.py

@@ -17,7 +17,7 @@ from ray.rllib.policy.policy_template import build_policy_class
 from ray.rllib.policy.sample_batch import SampleBatch
 from ray.rllib.utils.framework import try_import_torch
 from ray.rllib.utils.spaces.simplex import Simplex
-from ray.rllib.utils.torch_ops import apply_grad_clipping, \
+from ray.rllib.utils.torch_utils import apply_grad_clipping, \
     concat_multi_gpu_td_errors, huber_loss, l2_loss
 from ray.rllib.utils.typing import TrainerConfigDict, TensorType, \
     LocalOptimizer, GradInfoDict

rllib/agents/ddpg/tests/test_ddpg.py

@@ -14,7 +14,7 @@ from ray.rllib.utils.framework import try_import_tf, try_import_torch
 from ray.rllib.utils.numpy import fc, huber_loss, l2_loss, relu, sigmoid
 from ray.rllib.utils.test_utils import check, check_compute_single_action, \
     check_train_results, framework_iterator
-from ray.rllib.utils.torch_ops import convert_to_torch_tensor
+from ray.rllib.utils.torch_utils import convert_to_torch_tensor
 
 tf1, tf, tfv = try_import_tf()
 torch, _ = try_import_torch()

rllib/agents/dqn/dqn_torch_policy.py

@@ -19,7 +19,7 @@ from ray.rllib.policy.torch_policy import LearningRateSchedule
 from ray.rllib.utils.error import UnsupportedSpaceException
 from ray.rllib.utils.exploration.parameter_noise import ParameterNoise
 from ray.rllib.utils.framework import try_import_torch
-from ray.rllib.utils.torch_ops import apply_grad_clipping, \
+from ray.rllib.utils.torch_utils import apply_grad_clipping, \
     concat_multi_gpu_td_errors, FLOAT_MIN, huber_loss, \
     reduce_mean_ignore_inf, softmax_cross_entropy_with_logits
 from ray.rllib.utils.typing import TensorType, TrainerConfigDict

rllib/agents/dqn/r2d2_torch_policy.py

@@ -19,7 +19,7 @@ from ray.rllib.policy.policy_template import build_policy_class
 from ray.rllib.policy.sample_batch import SampleBatch
 from ray.rllib.policy.torch_policy import LearningRateSchedule
 from ray.rllib.utils.framework import try_import_torch
-from ray.rllib.utils.torch_ops import apply_grad_clipping, \
+from ray.rllib.utils.torch_utils import apply_grad_clipping, \
     concat_multi_gpu_td_errors, FLOAT_MIN, huber_loss, sequence_mask
 from ray.rllib.utils.typing import TensorType, TrainerConfigDict
 

rllib/agents/dqn/simple_q_torch_policy.py

@@ -16,7 +16,7 @@ from ray.rllib.policy.sample_batch import SampleBatch
 from ray.rllib.policy.torch_policy import TorchPolicy
 from ray.rllib.utils.annotations import override
 from ray.rllib.utils.framework import try_import_torch
-from ray.rllib.utils.torch_ops import concat_multi_gpu_td_errors, huber_loss
+from ray.rllib.utils.torch_utils import concat_multi_gpu_td_errors, huber_loss
 from ray.rllib.utils.typing import TensorType, TrainerConfigDict
 
 torch, nn = try_import_torch()

rllib/agents/dreamer/dreamer_torch_policy.py

@@ -11,7 +11,7 @@ from ray.rllib.policy.policy import Policy
 from ray.rllib.policy.policy_template import build_policy_class
 from ray.rllib.policy.sample_batch import SampleBatch
 from ray.rllib.utils.framework import try_import_torch
-from ray.rllib.utils.torch_ops import apply_grad_clipping
+from ray.rllib.utils.torch_utils import apply_grad_clipping
 from ray.rllib.utils.typing import AgentID
 
 torch, nn = try_import_torch()

rllib/agents/es/es.py

@@ -16,7 +16,7 @@ from ray.rllib.policy.sample_batch import DEFAULT_POLICY_ID
 from ray.rllib.utils import FilterManager
 from ray.rllib.utils.annotations import override
 from ray.rllib.utils.deprecation import Deprecated
-from ray.rllib.utils.torch_ops import set_torch_seed
+from ray.rllib.utils.torch_utils import set_torch_seed
 
 logger = logging.getLogger(__name__)
 

rllib/agents/es/es_torch_policy.py

@@ -13,7 +13,7 @@ from ray.rllib.utils.filter import get_filter
 from ray.rllib.utils.framework import try_import_torch
 from ray.rllib.utils.spaces.space_utils import get_base_struct_from_space, \
     unbatch
-from ray.rllib.utils.torch_ops import convert_to_torch_tensor
+from ray.rllib.utils.torch_utils import convert_to_torch_tensor
 
 torch, _ = try_import_torch()
 

rllib/agents/impala/vtrace_torch.py

@@ -34,7 +34,7 @@ from ray.rllib.agents.impala.vtrace_tf import VTraceFromLogitsReturns, \
 from ray.rllib.models.torch.torch_action_dist import TorchCategorical
 from ray.rllib.utils import force_list
 from ray.rllib.utils.framework import try_import_torch
-from ray.rllib.utils.torch_ops import convert_to_torch_tensor
+from ray.rllib.utils.torch_utils import convert_to_torch_tensor
 
 torch, nn = try_import_torch()
 

rllib/agents/impala/vtrace_torch_policy.py

@@ -12,7 +12,7 @@ from ray.rllib.policy.sample_batch import SampleBatch
 from ray.rllib.policy.torch_policy import LearningRateSchedule, \
     EntropyCoeffSchedule
 from ray.rllib.utils.framework import try_import_torch
-from ray.rllib.utils.torch_ops import apply_grad_clipping, \
+from ray.rllib.utils.torch_utils import apply_grad_clipping, \
     explained_variance, global_norm, sequence_mask
 
 torch, nn = try_import_torch()

rllib/agents/maml/maml_torch_policy.py

@@ -8,7 +8,7 @@ from ray.rllib.policy.sample_batch import SampleBatch
 from ray.rllib.agents.ppo.ppo_tf_policy import setup_config
 from ray.rllib.agents.ppo.ppo_torch_policy import vf_preds_fetches, \
     ValueNetworkMixin
-from ray.rllib.utils.torch_ops import apply_grad_clipping
+from ray.rllib.utils.torch_utils import apply_grad_clipping
 from ray.rllib.utils.framework import try_import_torch
 
 torch, nn = try_import_torch()

rllib/agents/marwil/marwil_torch_policy.py

@@ -8,7 +8,7 @@ from ray.rllib.evaluation.postprocessing import Postprocessing
 from ray.rllib.policy.policy_template import build_policy_class
 from ray.rllib.policy.sample_batch import SampleBatch
 from ray.rllib.utils.framework import try_import_torch
-from ray.rllib.utils.torch_ops import apply_grad_clipping, explained_variance
+from ray.rllib.utils.torch_utils import apply_grad_clipping, explained_variance
 from ray.rllib.utils.typing import TrainerConfigDict, TensorType
 from ray.rllib.policy.policy import Policy
 from ray.rllib.models.action_dist import ActionDistribution

rllib/agents/mbmpo/mbmpo.py

@@ -33,7 +33,7 @@ from ray.rllib.policy.sample_batch import DEFAULT_POLICY_ID, SampleBatch
 from ray.rllib.utils.deprecation import DEPRECATED_VALUE
 from ray.rllib.utils.metrics.learner_info import LEARNER_INFO
 from ray.rllib.utils.sgd import standardized
-from ray.rllib.utils.torch_ops import convert_to_torch_tensor
+from ray.rllib.utils.torch_utils import convert_to_torch_tensor
 from ray.rllib.utils.typing import EnvType, TrainerConfigDict
 from ray.util.iter import from_actors, LocalIterator
 

rllib/agents/mbmpo/mbmpo_torch_policy.py

@@ -16,7 +16,7 @@ from ray.rllib.policy.policy import Policy
 from ray.rllib.policy.policy_template import build_policy_class
 from ray.rllib.utils.error import UnsupportedSpaceException
 from ray.rllib.utils.framework import try_import_torch
-from ray.rllib.utils.torch_ops import apply_grad_clipping
+from ray.rllib.utils.torch_utils import apply_grad_clipping
 from ray.rllib.utils.typing import TrainerConfigDict
 
 torch, nn = try_import_torch()

rllib/agents/mbmpo/model_ensemble.py

@@ -7,7 +7,7 @@ from ray.rllib.evaluation.rollout_worker import get_global_worker
 from ray.rllib.policy.sample_batch import SampleBatch
 from ray.rllib.execution.common import STEPS_SAMPLED_COUNTER
 from ray.rllib.utils.typing import SampleBatchType
-from ray.rllib.utils.torch_ops import convert_to_torch_tensor
+from ray.rllib.utils.torch_utils import convert_to_torch_tensor
 
 torch, nn = try_import_torch()
 

rllib/agents/ppo/appo_torch_policy.py

@@ -28,8 +28,8 @@ from ray.rllib.policy.sample_batch import SampleBatch
 from ray.rllib.policy.torch_policy import EntropyCoeffSchedule, \
     LearningRateSchedule
 from ray.rllib.utils.framework import try_import_torch
-from ray.rllib.utils.torch_ops import apply_grad_clipping, explained_variance,\
-    global_norm, sequence_mask
+from ray.rllib.utils.torch_utils import apply_grad_clipping, \
+    explained_variance, global_norm, sequence_mask
 from ray.rllib.utils.typing import TensorType, TrainerConfigDict
 
 torch, nn = try_import_torch()

rllib/agents/ppo/ppo_torch_policy.py

@@ -17,7 +17,7 @@ from ray.rllib.policy.sample_batch import SampleBatch
 from ray.rllib.policy.torch_policy import EntropyCoeffSchedule, \
     LearningRateSchedule
 from ray.rllib.utils.framework import try_import_torch
-from ray.rllib.utils.torch_ops import apply_grad_clipping, \
+from ray.rllib.utils.torch_utils import apply_grad_clipping, \
     explained_variance, sequence_mask
 from ray.rllib.utils.typing import TensorType, TrainerConfigDict
 

rllib/agents/sac/rnnsac_torch_policy.py

@@ -13,7 +13,7 @@ from ray.rllib.models.torch.torch_action_dist import TorchDistributionWrapper
 from ray.rllib.policy.policy import Policy
 from ray.rllib.policy.sample_batch import SampleBatch
 from ray.rllib.utils.framework import try_import_torch
-from ray.rllib.utils.torch_ops import huber_loss, sequence_mask
+from ray.rllib.utils.torch_utils import huber_loss, sequence_mask
 from ray.rllib.utils.typing import \
     ModelInputDict, TensorType, TrainerConfigDict
 

rllib/agents/sac/sac_torch_policy.py

@@ -25,7 +25,7 @@ from ray.rllib.policy.torch_policy import TorchPolicy
 from ray.rllib.utils.annotations import override
 from ray.rllib.utils.framework import try_import_torch
 from ray.rllib.utils.spaces.simplex import Simplex
-from ray.rllib.utils.torch_ops import apply_grad_clipping, \
+from ray.rllib.utils.torch_utils import apply_grad_clipping, \
     concat_multi_gpu_td_errors, huber_loss
 from ray.rllib.utils.typing import LocalOptimizer, ModelInputDict, \
     TensorType, TrainerConfigDict

rllib/agents/sac/tests/test_sac.py

@@ -22,7 +22,7 @@ from ray.rllib.utils.numpy import fc, huber_loss, relu
 from ray.rllib.utils.spaces.simplex import Simplex
 from ray.rllib.utils.test_utils import check, check_compute_single_action, \
     check_train_results, framework_iterator
-from ray.rllib.utils.torch_ops import convert_to_torch_tensor
+from ray.rllib.utils.torch_utils import convert_to_torch_tensor
 from ray import tune
 
 tf1, tf, tfv = try_import_tf()

rllib/examples/centralized_critic.py

@@ -39,7 +39,7 @@ from ray.rllib.policy.torch_policy import LearningRateSchedule as TorchLR, \
 from ray.rllib.utils.framework import try_import_tf, try_import_torch
 from ray.rllib.utils.test_utils import check_learning_achieved
 from ray.rllib.utils.tf_utils import explained_variance, make_tf_callable
-from ray.rllib.utils.torch_ops import convert_to_torch_tensor
+from ray.rllib.utils.torch_utils import convert_to_torch_tensor
 
 tf1, tf, tfv = try_import_tf()
 torch, nn = try_import_torch()

rllib/examples/models/action_mask_model.py

@@ -5,7 +5,7 @@ from ray.rllib.models.tf.tf_modelv2 import TFModelV2
 from ray.rllib.models.torch.torch_modelv2 import TorchModelV2
 from ray.rllib.models.torch.fcnet import FullyConnectedNetwork as TorchFC
 from ray.rllib.utils.framework import try_import_tf, try_import_torch
-from ray.rllib.utils.torch_ops import FLOAT_MIN
+from ray.rllib.utils.torch_utils import FLOAT_MIN
 
 tf1, tf, tfv = try_import_tf()
 torch, nn = try_import_torch()

rllib/examples/models/parametric_actions_model.py

@@ -7,7 +7,7 @@ from ray.rllib.agents.dqn.dqn_torch_model import \
 from ray.rllib.models.tf.fcnet import FullyConnectedNetwork
 from ray.rllib.models.torch.fcnet import FullyConnectedNetwork as TorchFC
 from ray.rllib.utils.framework import try_import_tf, try_import_torch
-from ray.rllib.utils.torch_ops import FLOAT_MIN, FLOAT_MAX
+from ray.rllib.utils.torch_utils import FLOAT_MIN, FLOAT_MAX
 
 tf1, tf, tfv = try_import_tf()
 torch, nn = try_import_torch()

rllib/examples/models/trajectory_view_utilizing_models.py

@@ -4,7 +4,7 @@ from ray.rllib.models.torch.torch_modelv2 import TorchModelV2
 from ray.rllib.policy.view_requirement import ViewRequirement
 from ray.rllib.utils.framework import try_import_tf, try_import_torch
 from ray.rllib.utils.tf_utils import one_hot
-from ray.rllib.utils.torch_ops import one_hot as torch_one_hot
+from ray.rllib.utils.torch_utils import one_hot as torch_one_hot
 
 tf1, tf, tfv = try_import_tf()
 torch, nn = try_import_torch()

rllib/models/torch/attention_net.py

@@ -23,7 +23,7 @@ from ray.rllib.policy.sample_batch import SampleBatch
 from ray.rllib.policy.view_requirement import ViewRequirement
 from ray.rllib.utils.annotations import override
 from ray.rllib.utils.framework import try_import_torch
-from ray.rllib.utils.torch_ops import one_hot
+from ray.rllib.utils.torch_utils import one_hot
 from ray.rllib.utils.typing import ModelConfigDict, TensorType, List
 
 torch, nn = try_import_torch()

rllib/models/torch/complex_input_net.py

@@ -14,7 +14,7 @@ from ray.rllib.policy.sample_batch import SampleBatch
 from ray.rllib.utils.annotations import override
 from ray.rllib.utils.framework import try_import_torch
 from ray.rllib.utils.spaces.space_utils import flatten_space
-from ray.rllib.utils.torch_ops import one_hot
+from ray.rllib.utils.torch_utils import one_hot
 
 torch, nn = try_import_torch()
 

rllib/models/torch/modules/multi_head_attention.py

@@ -5,7 +5,7 @@
 """
 from ray.rllib.utils.framework import try_import_torch
 from ray.rllib.models.torch.misc import SlimFC
-from ray.rllib.utils.torch_ops import sequence_mask
+from ray.rllib.utils.torch_utils import sequence_mask
 from ray.rllib.utils.framework import TensorType
 
 torch, nn = try_import_torch()

rllib/models/torch/modules/relative_multi_head_attention.py

@@ -2,7 +2,7 @@ from typing import Union
 
 from ray.rllib.utils.framework import try_import_torch
 from ray.rllib.models.torch.misc import SlimFC
-from ray.rllib.utils.torch_ops import sequence_mask
+from ray.rllib.utils.torch_utils import sequence_mask
 from ray.rllib.utils.typing import TensorType
 
 torch, nn = try_import_torch()

rllib/models/torch/recurrent_net.py

@@ -11,7 +11,7 @@ from ray.rllib.policy.sample_batch import SampleBatch
 from ray.rllib.policy.view_requirement import ViewRequirement
 from ray.rllib.utils.annotations import override, DeveloperAPI
 from ray.rllib.utils.framework import try_import_torch
-from ray.rllib.utils.torch_ops import one_hot
+from ray.rllib.utils.torch_utils import one_hot
 from ray.rllib.utils.typing import ModelConfigDict, TensorType
 
 torch, nn = try_import_torch()

rllib/models/utils.py

@@ -28,7 +28,7 @@ def get_activation_fn(name: Optional[str] = None, framework: str = "tf"):
         if name in ["linear", None]:
             return None
         if name == "swish":
-            from ray.rllib.utils.torch_ops import Swish
+            from ray.rllib.utils.torch_utils import Swish
             return Swish
         _, nn = try_import_torch()
         if name == "relu":

rllib/policy/policy_template.py

@@ -14,7 +14,7 @@ from ray.rllib.utils import add_mixins, force_list, NullContextManager
 from ray.rllib.utils.annotations import override, DeveloperAPI
 from ray.rllib.utils.framework import try_import_torch, try_import_jax
 from ray.rllib.utils.metrics.learner_info import LEARNER_STATS_KEY
-from ray.rllib.utils.torch_ops import convert_to_non_torch_type
+from ray.rllib.utils.torch_utils import convert_to_non_torch_type
 from ray.rllib.utils.typing import ModelGradients, TensorType, \
     TrainerConfigDict
 

rllib/policy/torch_policy.py

@@ -26,7 +26,7 @@ from ray.rllib.utils.numpy import convert_to_numpy
 from ray.rllib.utils.schedules import PiecewiseSchedule
 from ray.rllib.utils.spaces.space_utils import normalize_action
 from ray.rllib.utils.threading import with_lock
-from ray.rllib.utils.torch_ops import convert_to_torch_tensor
+from ray.rllib.utils.torch_utils import convert_to_torch_tensor
 from ray.rllib.utils.typing import ModelGradients, ModelWeights, TensorType, \
     TensorStructType, TrainerConfigDict
 

rllib/utils/exploration/curiosity.py

@@ -18,7 +18,7 @@ from ray.rllib.utils.framework import try_import_tf, \
     try_import_torch
 from ray.rllib.utils.from_config import from_config
 from ray.rllib.utils.tf_utils import get_placeholder, one_hot as tf_one_hot
-from ray.rllib.utils.torch_ops import one_hot
+from ray.rllib.utils.torch_utils import one_hot
 from ray.rllib.utils.typing import FromConfigSpec, ModelConfigDict, TensorType
 
 tf1, tf, tfv = try_import_tf()

rllib/utils/exploration/epsilon_greedy.py

@@ -13,7 +13,7 @@ from ray.rllib.utils.framework import try_import_tf, try_import_torch, \
 from ray.rllib.utils.from_config import from_config
 from ray.rllib.utils.numpy import convert_to_numpy
 from ray.rllib.utils.schedules import Schedule, PiecewiseSchedule
-from ray.rllib.utils.torch_ops import FLOAT_MIN
+from ray.rllib.utils.torch_utils import FLOAT_MIN
 
 tf1, tf, tfv = try_import_tf()
 torch, _ = try_import_torch()

rllib/utils/framework.py

@@ -280,7 +280,7 @@ def get_activation_fn(name: Optional[str] = None, framework: str = "tf"):
         if name in ["linear", None]:
             return None
         if name in ["swish", "silu"]:
-            from ray.rllib.utils.torch_ops import Swish
+            from ray.rllib.utils.torch_utils import Swish
             return Swish
         _, nn = try_import_torch()
         if name == "relu":