fix: break shared-buffer memory leak in GatedDeltaNet cache (#1077)

Co-authored-by: Angelos Katharopoulos <[email protected]>

mlx_lm/models/kimi_linear.py

@@ -267,7 +267,7 @@ class ShortConv1d(nn.Module):
             positions = (ends[:, None] + mx.arange(n_keep))[..., None]
             new_state = mx.take_along_axis(conv_input, positions, axis=1)
         else:
-            new_state = conv_input[:, -n_keep:, :]
+            new_state = mx.contiguous(conv_input[:, -n_keep:, :])
 
         return out, new_state
 

mlx_lm/models/qwen3_5.py

@@ -157,7 +157,8 @@ class GatedDeltaNet(nn.Module):
             qkv = mx.where(mask[..., None], qkv, 0)
         conv_input = mx.concatenate([conv_state, qkv], axis=1)
         if cache is not None:
-            cache[0] = conv_input[:, -(self.conv_kernel_size - 1) :]
+            n_keep = self.conv_kernel_size - 1
+            cache[0] = mx.contiguous(conv_input[:, -n_keep:])
         conv_out = nn.silu(self.conv1d(conv_input))
 
         q, k, v = [

mlx_lm/models/qwen3_next.py

@@ -266,7 +266,7 @@ class Qwen3NextGatedDeltaNet(nn.Module):
                 positions = (ends[:, None] + mx.arange(n_keep))[..., None]
                 cache[0] = mx.take_along_axis(conv_input, positions, axis=1)
             else:
-                cache[0] = conv_input[:, -n_keep:, :]
+                cache[0] = mx.contiguous(conv_input[:, -n_keep:, :])
 
         conv_out = nn.silu(self.conv1d(conv_input))