diff --git a/README.md b/README.md
index d71b268..ff7dbf5 100644
--- a/README.md
+++ b/README.md
@@ -168,4 +168,4 @@ Runge-Kutta methods generally have less discretization error than standard diffu
   - To use the new solvers, select `adaptive_scipy` as the step size controller.
   - No fixed step size controllers are available for wrapped scipy methods. 
   - These methods do not support parallel IVP solve, meaning the batch elements are processed sequentially.
-  - Implicit solvers from `scipy.integrate` do not work for sampling diffusion models as the root finding step takes too long, so the implementation for them are skipped.
\ No newline at end of file
+  - Implicit solvers from `scipy.integrate` do not work for sampling diffusion models as the root finding step takes too long, so the implementation for them were skipped.
\ No newline at end of file
diff --git a/nodes/nodes_rk_sampler.py b/nodes/nodes_rk_sampler.py
index 513c916..30b8e6c 100644
--- a/nodes/nodes_rk_sampler.py
+++ b/nodes/nodes_rk_sampler.py
@@ -300,7 +300,7 @@ class RungeKuttaSamplerImpl:
             return self._call_torchode(model, x, sigmas, extra_args=extra_args, callback=callback, disable=disable)
         elif self.step_size_controller == "fixed_scheduled":
             if (self.method not in ADAPTIVE_METHODS) and (self.method not in FIXED_METHODS):
-                raise ValueError("fixed_scheduled only supports methods staring with `a` or `f`")
+                raise ValueError("fixed_scheduled only supports methods starting with `a` or `f`")
             return self._call_torchode(model, x, sigmas, extra_args=extra_args, callback=callback, disable=disable)
         elif self.step_size_controller == "adaptive_scipy":
             if self.method not in ADAPTIVE_SCIPY_METHODS: