diff --git a/kvpress/presses/decoding_press.py b/kvpress/presses/decoding_press.py
index 56d4b6b1..34eb1c40 100644
--- a/kvpress/presses/decoding_press.py
+++ b/kvpress/presses/decoding_press.py
@@ -3,10 +3,12 @@
 
 import logging
 from collections import defaultdict
+from contextlib import contextmanager
 from dataclasses import dataclass
 
 import torch
 import torch.nn as nn
+from transformers import PreTrainedModel
 from transformers.cache_utils import QuantizedCache
 
 from kvpress.presses.adakv_press import AdaKVPress
@@ -179,6 +181,14 @@ def reset(self):
         self.hidden_states_buffer = defaultdict(list)
         self.layer_step_counts = defaultdict(int)
 
+    @contextmanager
+    def __call__(self, model: PreTrainedModel):
+        try:
+            with super().__call__(model):
+                yield
+        finally:
+            self.reset()
+
     def _find_target_compression_ratio(self, q_len: int, target_tokens: int) -> float:
         """
         Find the compression ratio that results in exactly target_tokens after int() rounding.