diff --git a/dexteroushandenvs/algorithms/td3/storage.py b/dexteroushandenvs/algorithms/td3/storage.py
deleted file mode 100644
index a13a7e4..0000000
--- a/dexteroushandenvs/algorithms/td3/storage.py
+++ /dev/null
@@ -1,61 +0,0 @@
-import torch
-from torch.utils.data.sampler import BatchSampler, SequentialSampler, SubsetRandomSampler
-
-
-class ReplayBuffer:
-
-    def __init__(self, num_envs, num_transitions_per_env, obs_shape, states_shape, actions_shape, device='cpu', sampler='sequential'):
-
-        self.device = device
-        self.sampler = sampler
-
-        # Core
-        self.observations = torch.zeros(num_transitions_per_env, num_envs, *obs_shape, device=self.device)
-        self.states = torch.zeros(num_transitions_per_env, num_envs, *states_shape, device=self.device)
-        self.rewards = torch.zeros(num_transitions_per_env, num_envs, 1, device=self.device)
-        self.next_observations = torch.zeros(num_transitions_per_env, num_envs, *obs_shape, device=self.device)
-        self.actions = torch.zeros(num_transitions_per_env, num_envs, *actions_shape, device=self.device)
-        self.dones = torch.zeros(num_transitions_per_env, num_envs, 1, device=self.device).byte()
-
-        self.num_transitions_per_env = num_transitions_per_env
-        self.num_envs = num_envs
-
-        self.step = 0
-
-    def add_transitions(self, observations, states, actions, rewards, next_obs ,dones):
-        if self.step >= self.num_transitions_per_env:
-            #TODO: 有点bug 清不掉0 后续改下
-            self.step = (self.step + 1) % self.num_transitions_per_env
-            # raise AssertionError("Rollout buffer overflow")
-
-        self.observations[self.step].copy_(observations)
-        self.states[self.step].copy_(states)
-        self.actions[self.step].copy_(actions)
-        self.rewards[self.step].copy_(rewards.view(-1, 1))
-        self.next_observations[self.step].copy_(next_obs)
-        self.dones[self.step].copy_(dones.view(-1, 1))
-
-        self.step += 1
-
-
-    def get_statistics(self):
-        done = self.dones.cpu()
-        done[-1] = 1
-        flat_dones = done.permute(1, 0, 2).reshape(-1, 1)
-        done_indices = torch.cat((flat_dones.new_tensor([-1], dtype=torch.int64), flat_dones.nonzero(as_tuple=False)[:, 0]))
-        trajectory_lengths = (done_indices[1:] - done_indices[:-1])
-        return trajectory_lengths.float().mean(), self.rewards.mean()
-
-    def mini_batch_generator(self, num_mini_batches):
-        batch_size = self.num_envs * self.num_transitions_per_env
-        mini_batch_size = batch_size // num_mini_batches
-
-        if self.sampler == "sequential":
-            # For physics-based RL, each environment is already randomized. There is no value to doing random sampling
-            # but a lot of CPU overhead during the PPO process. So, we can just switch to a sequential sampler instead
-            subset = SequentialSampler(range(batch_size))
-        elif self.sampler == "random":
-            subset = SubsetRandomSampler(range(batch_size))
-
-        batch = BatchSampler(subset, mini_batch_size, drop_last=True)
-        return batch
\ No newline at end of file