agilerl 2.4.1.dev3__py3-none-any.whl → 2.4.2__py3-none-any.whl

agilerl/algorithms/bc_lm.py

@@ -55,7 +55,7 @@ class BC_LM(nn.Module):
         prefix_embs: Optional[torch.Tensor] = None,
         prefix_attn_mask: Optional[torch.Tensor] = None,
         remove_prefix_position_embs: bool = False,
-        **kwargs
+        **kwargs,
     ):
         # tokens – b,t
         # attn_mask – b,t
@@ -83,7 +83,7 @@ class BC_LM(nn.Module):
             tok_emb=input_embeddings,
             attn_mask=input_attn_mask,
             pos=position_ids,
-            **kwargs
+            **kwargs,
         )
         return model_outputs, model_past_key_values
 

agilerl/algorithms/core/base.py

@@ -652,6 +652,9 @@ class EvolvableAlgorithm(ABC, metaclass=RegistryMeta):
         :type training: bool
         """
         self.training = training
+        for name, network in self.evolvable_attributes(networks_only=True).items():
+            if "actor" in name:
+                network.train(mode=training)
 
     def get_lr_names(self) -> list[str]:
         """Returns the learning rates of the algorithm."""
@@ -2063,8 +2066,7 @@ class LLMAlgorithm(EvolvableAlgorithm, ABC):
         accelerator: Optional[Accelerator] = None,
     ) -> None:
         raise NotImplementedError(
-            "The load class method is not supported for this algorithm class."
-            """
+            "The load class method is not supported for this algorithm class." """
             To load a saved LLM, please load the model as follows, and then re-instantiate the GRPO
             class, using the pre-trained model.
 

agilerl/algorithms/dqn.py

@@ -7,7 +7,6 @@ import torch._dynamo
 import torch.nn as nn
 import torch.optim as optim
 from gymnasium import spaces
-from tensordict import TensorDict, from_module
 from tensordict.nn import CudaGraphModule
 
 from agilerl.algorithms.core import OptimizerWrapper, RLAlgorithm
@@ -144,8 +143,8 @@ class DQN(RLAlgorithm):
             self.actor = create_actor()
             self.actor_target = create_actor()
 
-        # Copy over actor weights to target
-        self.init_hook()
+        # Initialize target network (same pattern as DDPG; post-mutation sync via reinit_shared_networks)
+        self.actor_target.load_state_dict(self.actor.state_dict())
 
         # Initialize optimizer with OptimizerWrapper
         self.optimizer = OptimizerWrapper(
@@ -172,7 +171,7 @@ class DQN(RLAlgorithm):
             self.update = CudaGraphModule(self.update)
             self._get_action = CudaGraphModule(self._get_action)
 
-        # Register DQN network groups and mutation hook
+        # Register DQN network groups
         self.register_network_group(
             NetworkGroup(
                 eval_network=self.actor,
@@ -180,27 +179,6 @@ class DQN(RLAlgorithm):
                 policy=True,
             )
         )
-        self.register_mutation_hook(self.init_hook)
-
-    def init_hook(self) -> None:
-        """Resets module parameters for the detached and target networks."""
-        param_vals: TensorDict = from_module(self.actor).detach()
-
-        # NOTE: This removes the target params from the computation graph which
-        # reduces memory overhead and speeds up training, however these won't
-        # appear in the modules parameters
-        target_params: TensorDict = param_vals.clone().lock_()
-
-        # This hook is prompted after performing architecture mutations on policy / evaluation
-        # networks, which will fail since the target network is a shared network that won't be
-        # reintiialized until the end. We can bypass the error safely for this reason.
-        try:
-            target_params.to_module(self.actor_target)
-        except KeyError:
-            pass
-        finally:
-            self.param_vals = param_vals
-            self.target_params = target_params
 
     def get_action(
         self,
@@ -260,8 +238,10 @@ class DQN(RLAlgorithm):
         :return: Selected action(s) as tensor
         :rtype: torch.Tensor
         """
+        self.actor.eval()
         with torch.no_grad():
             q_values = self.actor(obs)
+        self.actor.train()
 
         # Masked random actions
         masked_random_values = torch.rand_like(q_values) * action_mask

agilerl/algorithms/ilql.py

@@ -331,7 +331,7 @@ class ILQL(nn.Module):
             tok_emb=input_embeddings,
             attn_mask=input_attn_mask,
             pos=position_ids,
-            **qv_kwargs
+            **qv_kwargs,
         )
         hidden_states = model_hidden_states[-1][:, prefix_t:, :]
 
@@ -345,7 +345,7 @@ class ILQL(nn.Module):
                 tok_emb=target_input_embeddings,
                 attn_mask=input_attn_mask,
                 pos=position_ids,
-                **target_kwargs
+                **target_kwargs,
             )
         target_hidden_states = target_hidden_states[-1][:, prefix_t:, :]
 
@@ -373,7 +373,7 @@ class ILQL(nn.Module):
                         tok_emb=policy_input_embeddings,
                         attn_mask=input_attn_mask,
                         pos=position_ids,
-                        **policy_kwargs
+                        **policy_kwargs,
                     )
             else:
                 (
@@ -385,7 +385,7 @@ class ILQL(nn.Module):
                     tok_emb=policy_input_embeddings,
                     attn_mask=input_attn_mask,
                     pos=position_ids,
-                    **policy_kwargs
+                    **policy_kwargs,
                 )
             policy_hidden_states = policy_hidden_states[-1][:, prefix_t:, :]
 
@@ -626,7 +626,7 @@ class ILQL(nn.Module):
         qv_kwargs=None,
         policy_kwargs=None,
         target_kwargs=None,
-        **kwargs
+        **kwargs,
     ):
         prepared_inputs = self.prepare_inputs(items)
         tokens, attn_mask = prepared_inputs["tokens"], prepared_inputs["attn_mask"]

agilerl/algorithms/ippo.py

@@ -671,7 +671,7 @@ class IPPO(MultiAgentRLAlgorithm):
         :param action_space: Action space for the agent
         :type action_space: gymnasium.spaces
         """
-        (states, actions, log_probs, rewards, dones, values, next_state, next_done) = (
+        states, actions, log_probs, rewards, dones, values, next_state, next_done = (
             experiences
         )
 

agilerl/networks/q_networks.py

@@ -248,7 +248,7 @@ class RainbowQNetwork(EvolvableNetwork):
             num_atoms=self.num_atoms,
             support=self.support,
             device=self.device,
-            **net_config
+            **net_config,
         )
 
     def forward(
@@ -279,7 +279,7 @@ class RainbowQNetwork(EvolvableNetwork):
             num_atoms=self.num_atoms,
             support=self.support,
             device=self.device,
-            **self.head_net.net_config
+            **self.head_net.net_config,
         )
 
         self.head_net = EvolvableModule.preserve_parameters(self.head_net, head_net)

agilerl/protocols.py

@@ -385,7 +385,7 @@ class PreTrainedModelProtocol(Protocol):
         input_ids: torch.Tensor,
         attention_mask: Optional[torch.Tensor] = None,
         generation_config: Optional["GenerationConfigProtocol"] = None,
-        **kwargs: Any
+        **kwargs: Any,
     ) -> torch.Tensor: ...
     def forward(self, *args: Any, **kwargs: Any) -> Any: ...
     def parameters(self) -> Generator: ...
@@ -416,7 +416,7 @@ class PeftModelProtocol(Protocol):
         input_ids: torch.Tensor,
         attention_mask: Optional[torch.Tensor] = None,
         generation_config: Optional["GenerationConfigProtocol"] = None,
-        **kwargs: Any
+        **kwargs: Any,
     ) -> torch.Tensor: ...
     def forward(self, *args: Any, **kwargs: Any) -> Any: ...
     def parameters(self) -> Generator: ...

agilerl/training/train_off_policy.py

@@ -145,7 +145,7 @@ def train_off_policy(
     assert isinstance(max_steps, int), "Number of steps must be an integer."
     assert isinstance(evo_steps, int), "Evolution frequency must be an integer."
     assert isinstance(eps_start, float), "Starting epsilon must be a float."
-    assert isinstance(eps_end, float), "Final value of epsilone must be a float."
+    assert isinstance(eps_end, float), "Final value of epsilon must be a float."
     assert isinstance(eps_decay, float), "Epsilon decay rate must be a float."
     if target is not None:
         assert isinstance(

agilerl/utils/algo_utils.py

@@ -6,7 +6,7 @@ from collections import OrderedDict, defaultdict
 from dataclasses import dataclass
 from functools import singledispatch
 from numbers import Number
-from typing import Any, Optional, Union
+from typing import Any, ForwardRef, Optional, Union
 
 import numpy as np
 import torch
@@ -50,7 +50,7 @@ if HAS_LLM_DEPENDENCIES:
 
     PreTrainedModelType = Union[PeftModel, PreTrainedModel]
 else:
-    PreTrainedModelType = Union["PeftModel", "PreTrainedModel"]
+    PreTrainedModelType = Union[ForwardRef("PeftModel"), ForwardRef("PreTrainedModel")]
 
 
 def check_supported_space(observation_space: GymSpaceType) -> None:

agilerl/utils/llm_utils.py

@@ -105,7 +105,6 @@ class HuggingFaceGym(gym.Env, ABC):
         accelerator: Accelerator | None = None,
         seed: int = 42,
     ) -> None:
-
         self.name = train_dataset.info.dataset_name
         self.tokenizer = tokenizer
         self.data_batch_size_per_gpu = data_batch_size_per_gpu
@@ -431,7 +430,6 @@ class ReasoningGym(HuggingFaceGym):
         """
 
         def collate_fn(batch):
-
             questions = [item["question"] for item in batch]
             answers = [item["answer"] for item in batch]
 
@@ -551,7 +549,6 @@ class PreferenceGym(HuggingFaceGym):
         """
 
         def collate_fn(batch: list[dict[str, str]]) -> dict[str, str]:
-
             prompts = [item["prompt"] for item in batch]
             chosen = [item["chosen"] for item in batch]
             rejected = [item["rejected"] for item in batch]

agilerl/wrappers/agent.py

@@ -597,8 +597,8 @@ class AsyncAgentsWrapper(AgentWrapper[MultiAgentRLAlgorithm]):
         :return: Learning information
         :rtype: Any
         """
-        (states, actions, log_probs, rewards, dones, values, next_state, next_done) = (
-            map(self.stack_experiences, experiences)
+        states, actions, log_probs, rewards, dones, values, next_state, next_done = map(
+            self.stack_experiences, experiences
         )
 
         # Handle case where we haven't collected a next state for each sub-agent

{agilerl-2.4.1.dev3.dist-info → agilerl-2.4.2.dist-info}/METADATA

@@ -1,52 +1,51 @@
 Metadata-Version: 2.4
 Name: agilerl
-Version: 2.4.1.dev3
+Version: 2.4.2
 Summary: AgileRL is a deep reinforcement learning library focused on improving RL development through RLOps.
-License: Apache 2.0
+Author-email: Nick Ustaran-Anderegg <dev@agilerl.com>
+License-Expression: Apache-2.0
 License-File: LICENSE
-Author: Nick Ustaran-Anderegg
-Author-email: dev@agilerl.com
-Requires-Python: >=3.10,<3.13
-Classifier: License :: Other/Proprietary License
-Classifier: Programming Language :: Python :: 3
-Classifier: Programming Language :: Python :: 3.10
-Classifier: Programming Language :: Python :: 3.11
-Classifier: Programming Language :: Python :: 3.12
+Requires-Python: <3.13,>=3.10
+Requires-Dist: accelerate~=1.7.0
+Requires-Dist: dill~=0.3.7
+Requires-Dist: fastrand~=1.3.0
+Requires-Dist: flatten-dict~=0.4.2
+Requires-Dist: google-cloud-storage~=2.5.0
+Requires-Dist: gymnasium~=1.0.0
+Requires-Dist: h5py~=3.8.0
+Requires-Dist: hydra-core~=1.3.2
+Requires-Dist: jax[cpu]~=0.4.31
+Requires-Dist: matplotlib<3.10,~=3.9.4
+Requires-Dist: minari[all]==0.5.2
+Requires-Dist: numpy~=1.26.4
+Requires-Dist: omegaconf~=2.3.0
+Requires-Dist: packaging>=20.0
+Requires-Dist: pandas~=2.2.3
+Requires-Dist: pettingzoo~=1.23.1
+Requires-Dist: pre-commit~=3.4.0
+Requires-Dist: pygame~=2.6.0
+Requires-Dist: pymunk~=6.2.0
+Requires-Dist: redis~=4.4.4
+Requires-Dist: scipy~=1.12.0
+Requires-Dist: supersuit~=3.9.0
+Requires-Dist: tensordict~=0.8
+Requires-Dist: termcolor~=1.1.0
+Requires-Dist: torch==2.7.1
+Requires-Dist: tqdm~=4.66.4
+Requires-Dist: ucimlrepo~=0.0.3
+Requires-Dist: wandb~=0.17.6
 Provides-Extra: all
+Requires-Dist: datasets==4.4.1; extra == 'all'
+Requires-Dist: deepspeed~=0.17.1; extra == 'all'
+Requires-Dist: peft~=0.18.0; extra == 'all'
+Requires-Dist: transformers~=4.57.1; extra == 'all'
+Requires-Dist: vllm~=0.10.0; extra == 'all'
 Provides-Extra: llm
-Requires-Dist: SuperSuit (>=3.9.0,<4.0.0)
-Requires-Dist: accelerate (>=1.7.0,<2.0.0)
-Requires-Dist: datasets (==4.4.1) ; extra == "llm" or extra == "all"
-Requires-Dist: deepspeed (>=0.17.1,<0.18.0) ; extra == "llm" or extra == "all"
-Requires-Dist: dill (>=0.3.7,<0.4.0)
-Requires-Dist: fastrand (>=1.3.0,<2.0.0)
-Requires-Dist: flatten_dict (>=0.4.2,<0.5.0)
-Requires-Dist: google-cloud-storage (>=2.5.0,<3.0.0)
-Requires-Dist: gymnasium (>=1.0.0,<2.0.0)
-Requires-Dist: h5py (>=3.8.0,<4.0.0)
-Requires-Dist: hydra-core (>=1.3.2,<2.0.0)
-Requires-Dist: jax[cpu] (>=0.4.31,<0.5.0)
-Requires-Dist: matplotlib (>=3.9.4,<3.10.0)
-Requires-Dist: minari[all] (==0.5.2)
-Requires-Dist: numpy (>=1.26.4,<2.0.0)
-Requires-Dist: omegaconf (>=2.3.0,<3.0.0)
-Requires-Dist: packaging (>=20.0)
-Requires-Dist: pandas (>=2.2.3,<3.0.0)
-Requires-Dist: peft (>=0.18.0,<0.19.0) ; extra == "llm" or extra == "all"
-Requires-Dist: pettingzoo (>=1.23.1,<2.0.0)
-Requires-Dist: pre-commit (>=3.4.0,<4.0.0)
-Requires-Dist: pygame (>=2.6.0,<3.0.0)
-Requires-Dist: pymunk (>=6.2.0,<7.0.0)
-Requires-Dist: redis (>=4.4.4,<5.0.0)
-Requires-Dist: scipy (>=1.12.0,<2.0.0)
-Requires-Dist: tensordict (>=0.8,<0.9)
-Requires-Dist: termcolor (>=1.1.0,<2.0.0)
-Requires-Dist: torch (==2.7.1)
-Requires-Dist: tqdm (>=4.66.4,<5.0.0)
-Requires-Dist: transformers (>=4.57.1,<5.0.0) ; extra == "llm" or extra == "all"
-Requires-Dist: ucimlrepo (>=0.0.3,<0.0.4)
-Requires-Dist: vllm (==0.10.0) ; extra == "llm" or extra == "all"
-Requires-Dist: wandb (>=0.17.6,<0.18.0)
+Requires-Dist: datasets==4.4.1; extra == 'llm'
+Requires-Dist: deepspeed~=0.17.1; extra == 'llm'
+Requires-Dist: peft~=0.18.0; extra == 'llm'
+Requires-Dist: transformers~=4.57.1; extra == 'llm'
+Requires-Dist: vllm~=0.10.0; extra == 'llm'
 Description-Content-Type: text/markdown
 
 <p align="center">
@@ -363,4 +362,3 @@ title = {{AgileRL}},
 url = {https://github.com/AgileRL/AgileRL}
 }
 ```
-

{agilerl-2.4.1.dev3.dist-info → agilerl-2.4.2.dist-info}/RECORD

@@ -1,24 +1,26 @@
 agilerl/__init__.py,sha256=0hZjnAULURFWpshG_mhNdaHhf8nlc7h2sR7CLEqup54,572
+agilerl/protocols.py,sha256=AEOOsCc4zbYWqAfuZDb1Eki0Cu3QLTB42NU3-kNDZXI,14054
+agilerl/typing.py,sha256=JtLhZMNyFzrnSeos6ltWyD_8yWFkc8Zx-OIC3d1CPQc,5442
 agilerl/algorithms/__init__.py,sha256=5N4DqCEETuFBlhnzf7XEQzIClRXX9e-FxQqQHgLh3Es,661
-agilerl/algorithms/bc_lm.py,sha256=dDCN--Y49wJA_msVB_r8XYgLYXSYeJItYyhSD41bFFk,22946
-agilerl/algorithms/core/__init__.py,sha256=kKGnzj4TGRZKk2J6jcaKkK3s1LjCYu979o8u8OJUZjI,268
-agilerl/algorithms/core/base.py,sha256=LeFN0l17oCUxp23zFayq8tr9RFbSw--68TPa1FwobuA,121970
-agilerl/algorithms/core/optimizer_wrapper.py,sha256=UQTlnv-mbNGlQ3RX9ocHtczXhTZq1MBKO6OdoQ879uM,13086
-agilerl/algorithms/core/registry.py,sha256=ndaw9U814tHrPBhEPO9kLIDNKmLStTwLXPsnu-nnj8c,19991
+agilerl/algorithms/bc_lm.py,sha256=aL1ibo8Itv--A4yaW5I55fbx6sKRtRqqTZw3UDTAl-s,22948
 agilerl/algorithms/cqn.py,sha256=3zE6LPWPV8ut5hLPllw3yhY_amonbiSmbBXJU0-7Zo4,12583
 agilerl/algorithms/ddpg.py,sha256=uau1E37D9SARlf_bTswfZQGQRobh9tOcB6hoRpszx_g,21365
 agilerl/algorithms/dpo.py,sha256=kN2wp2Ms_2sFiJcmqpVPxG4XHoJis6l6BQlSCsj07pk,15777
-agilerl/algorithms/dqn.py,sha256=P05AspMruXghyqWGzXj4t0x6m6Pl9MKt8EKh3RP2yBU,17105
+agilerl/algorithms/dqn.py,sha256=3WYga_sVDflP1xVUJ2u-24jcmC_a5F0EXfthDJ5fbpQ,16210
 agilerl/algorithms/dqn_rainbow.py,sha256=HyP-jkiVOkBUJmvpUlrB6VHo8m-AO2Z84M3Zb_ZP6fQ,20483
 agilerl/algorithms/grpo.py,sha256=9VvRf4jQNDOfUlkKDZBNiiBACUybgeOxSQgnszjm2BM,19237
-agilerl/algorithms/ilql.py,sha256=vX070xfPFxNKWh6oEc_LERUJx80JQq8oMzZ8ESBOUgE,79844
-agilerl/algorithms/ippo.py,sha256=W9FDLf5bznG-RvfJs8Gqpa2ARGReqmPB9xW9mu2Mj-c,39085
+agilerl/algorithms/ilql.py,sha256=yQ6v6Y7n4JtsknCyhXOoJWMu-jbZX8CsLoitsEG2_YY,79849
+agilerl/algorithms/ippo.py,sha256=2JBPYnXGBxVbgkvy5BEa_m3Y4knKuIMA0EFNR3YADsQ,39083
 agilerl/algorithms/maddpg.py,sha256=qVXDyb_W51lZtvst4K3yiosSy58BEBYbck8wF8CViBA,33908
 agilerl/algorithms/matd3.py,sha256=n17y6PvM51r290Def_QeFT4p7TMo54MIDLN30XqlMk8,37926
 agilerl/algorithms/neural_ts_bandit.py,sha256=jL_5mnExjMZdiIdwMWXT1XH-hWtaIiSokxi_n_qGTDY,11790
 agilerl/algorithms/neural_ucb_bandit.py,sha256=wwo2sUNkIFtDDEOHIOp9aWhf5oeO9goi9p48tdH1Uno,11960
 agilerl/algorithms/ppo.py,sha256=yAkgZT7WbZKn2oq62DFDPcfAmnRHomVPm4yNlI9-B-c,53025
 agilerl/algorithms/td3.py,sha256=gFYlwwxYQgaWGDT5a-c3AOwI5WGQv4J4eeBotw1-fZY,23017
+agilerl/algorithms/core/__init__.py,sha256=kKGnzj4TGRZKk2J6jcaKkK3s1LjCYu979o8u8OJUZjI,268
+agilerl/algorithms/core/base.py,sha256=R0GyAIC3CaRzNKkufU_omfKzpasGefxCFNq5yvvwQ78,122119
+agilerl/algorithms/core/optimizer_wrapper.py,sha256=UQTlnv-mbNGlQ3RX9ocHtczXhTZq1MBKO6OdoQ879uM,13086
+agilerl/algorithms/core/registry.py,sha256=ndaw9U814tHrPBhEPO9kLIDNKmLStTwLXPsnu-nnj8c,19991
 agilerl/components/__init__.py,sha256=cc3bYeOdsNp-Puj_4_Ukj3kwmEqUqFeUo-5dZ3tP47o,292
 agilerl/components/data.py,sha256=KiXS4OPgC0VpM9cP8HMDoDhxYX-khL9vKEi2pYIWd7E,3832
 agilerl/components/multi_agent_replay_buffer.py,sha256=VfT90DhlrgMomzW_8Nw5zQrD4908hFLMgg0kXpy1ZHE,8604
@@ -53,9 +55,8 @@ agilerl/networks/base.py,sha256=Lkhj0yujVgNDKXKn_ea0hBFDSyStDwL5AuEyFnyGjmE,2163
 agilerl/networks/custom_modules.py,sha256=n6WR5DsXBQwQtvQk6lHiXP-DR-Ma6lGjrOtySSrIAiA,6843
 agilerl/networks/distributions.py,sha256=mzntWgwoEdZKAspInbmvfc6_0rGuPdquqQyQkVSWvoo,18252
 agilerl/networks/distributions_experimental.py,sha256=K6_EYflAlR6qRouRr6SJXnT19w7QhOA1bwN7kCl3DJ8,18890
-agilerl/networks/q_networks.py,sha256=a1Arze6GypKprxUQObbpJQbikmY5LtrvAAnEyoTrcLM,17284
+agilerl/networks/q_networks.py,sha256=pgX7lg-_bo724A7BC-b0ViLgXhWA5xbvoLnRnn8f9sU,17286
 agilerl/networks/value_networks.py,sha256=ZLX5vQIxeV65uxOzv2r5QMxF_-fzFT8N1et3lHdQP7E,4630
-agilerl/protocols.py,sha256=SQ8T79jmZAqlm2fJ1Qo0kefU5w2c4Mh_wUk9RtiPego,14052
 agilerl/rollouts/__init__.py,sha256=dGR9BnXliQI6yvXPwecV7g5TCtCEPbyIB-W1a5evBBY,130
 agilerl/rollouts/on_policy.py,sha256=VOxUjwzyYngzrTEW9asXsgz1O6lRTUn_PijmjqtzGwQ,8036
 agilerl/training/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -63,16 +64,15 @@ agilerl/training/train_bandits.py,sha256=pi6GFQrGsFkqgD8V69ayVlzcNUPMIF3PYaBEgJU
 agilerl/training/train_llm.py,sha256=1KxiJQGPLCVxMqsVNUWzsGHEwDL9ehvQ7a3gEELr2zM,27602
 agilerl/training/train_multi_agent_off_policy.py,sha256=p1VOBDqyt14LD5HUQ-YF5m2jce_LphgYa38DP4asY30,23349
 agilerl/training/train_multi_agent_on_policy.py,sha256=WDtUTpIpPuQpPdZN-1H_gwqHICyPRLfWIJeyYtClQKc,24427
-agilerl/training/train_off_policy.py,sha256=zWOPEPo8l6mAZw3XrYQousvv5tqou7c11TMlKJ8cqd0,23512
+agilerl/training/train_off_policy.py,sha256=iyMHnFrOjjuPxcIesrg9WFRmDFxXXI1guqeXMVb5XXg,23511
 agilerl/training/train_offline.py,sha256=qAlr3lGQf7EfSSmTtmohi80rUN4HMha955q3pae6TCY,13406
 agilerl/training/train_on_policy.py,sha256=iQEIHq_JgBIBH2GPJeLN6QmPRho-_beUdro1H9DPkUA,19360
-agilerl/typing.py,sha256=JtLhZMNyFzrnSeos6ltWyD_8yWFkc8Zx-OIC3d1CPQc,5442
 agilerl/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-agilerl/utils/algo_utils.py,sha256=Ue9uR5R_QywZbO7jvnQPTVAn6STLT9f-_nwrygs4Iz4,60376
+agilerl/utils/algo_utils.py,sha256=tFH89AD63qsN80xeyn9m85qRfjD12VN0nCan2_qRjT4,60412
 agilerl/utils/cache.py,sha256=8Q1SYbTxQYzIn40UMy32EWMvtgaduY1k5jqwPihxJ_Q,3418
 agilerl/utils/evolvable_networks.py,sha256=cIJHzadFOaK0aAqwn96HvnuH4atLBxrQ3cwpR1nxvUo,23265
 agilerl/utils/ilql_utils.py,sha256=dU_vbwOB6VsODGGu_hOyDN_xRtFKVhZbxMISFlAUM5s,2293
-agilerl/utils/llm_utils.py,sha256=rc4fnqw3z1RvKdDUisX4THbRTkAWeg84SPt7VTd_hJY,26594
+agilerl/utils/llm_utils.py,sha256=Rdwfo3L3TDyqfk4QfHrRRZ4-r8nblvMXp3-Qnf9W5k8,26591
 agilerl/utils/log_utils.py,sha256=OIhj86V97-ijlUENic2WKIWipB5ITJyBIGM_ZPZg5Vo,4401
 agilerl/utils/minari_utils.py,sha256=WNFzt9ZQuvWy3w84MFhhGkA0e9MAgc4KSI_cmPgFTBo,5109
 agilerl/utils/probe_envs.py,sha256=q2uyPQW7mbo9x4c_Yq9vi2Yu1X9qyLm43adET9SFf9Y,39796
@@ -84,12 +84,12 @@ agilerl/vector/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 agilerl/vector/pz_async_vec_env.py,sha256=uj9TyCn0SWksTUOW84RGspMkXqdGG-wjr86w08uCMb0,36742
 agilerl/vector/pz_vec_env.py,sha256=sFVqm8eecxVHahTpFZEE3fvyZrmp2vMu0GECik8el6M,5978
 agilerl/wrappers/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-agilerl/wrappers/agent.py,sha256=fjwY8JSzUBQDuXtgEsl8d2K9QxlmULobERjwq3WOyzc,23162
+agilerl/wrappers/agent.py,sha256=VyWju26YIfN4g8NCxTM-zMc3IGf5YKCXtyoeHynfEXE,23158
 agilerl/wrappers/learning.py,sha256=nSVMg6eUBWn13NNdIFgCEHj31CaN_dGryQa13SmMvBw,2774
 agilerl/wrappers/make_evolvable.py,sha256=sb9oAorGAayrD_6lNbyvHhefA_RKO4bSSNjqS6u9UhI,51079
 agilerl/wrappers/pettingzoo_wrappers.py,sha256=Pw8VzabxfYCw5ad15y5J3rAH1teA6nVVo0RHCTTdOPQ,2063
 agilerl/wrappers/utils.py,sha256=pENFH2AxsXd22s8HGUeM-jRowC0tmjHLWjqDwIq12l8,2194
-agilerl-2.4.1.dev3.dist-info/METADATA,sha256=ahIiSFnYkAUr_Dwia-i2KDuUcJm30WOGaAIaFrGNB30,20565
-agilerl-2.4.1.dev3.dist-info/WHEEL,sha256=zp0Cn7JsFoX2ATtOhtaFYIiE2rmFAD4OcMhtUki8W3U,88
-agilerl-2.4.1.dev3.dist-info/licenses/LICENSE,sha256=vPX_VnIseflXJ30mQvwbXZoe208EtIr9ZVrl6cfdQNs,11720
-agilerl-2.4.1.dev3.dist-info/RECORD,,
+agilerl-2.4.2.dist-info/METADATA,sha256=ohWA3cJL3JmVZr5svRWRslLWbCX5Lbx46ceqL9MWwgQ,20164
+agilerl-2.4.2.dist-info/WHEEL,sha256=WLgqFyCfm_KASv4WHyYy0P3pM_m7J5L9k2skdKLirC8,87
+agilerl-2.4.2.dist-info/licenses/LICENSE,sha256=vPX_VnIseflXJ30mQvwbXZoe208EtIr9ZVrl6cfdQNs,11720
+agilerl-2.4.2.dist-info/RECORD,,

{agilerl-2.4.1.dev3.dist-info → agilerl-2.4.2.dist-info}/WHEEL

@@ -1,4 +1,4 @@
 Wheel-Version: 1.0
-Generator: poetry-core 2.2.1
+Generator: hatchling 1.28.0
 Root-Is-Purelib: true
 Tag: py3-none-any