Is Instantaneous axis of rotation of a disk undergoing pure rolling an inertial frame? My teacher told that instantaneous axis of rotation of a disk in pure rolling that is point of contact of disk with ground is an non inertial frame due to radial acceleration. But I am confused that why would Instantaneous axis of rotation ( Point of contact of disk with ground in pure rolling) whose velocity is zero have an radial acceleration as radial acceleration = v²/ R. Another argument to support my thinking is that let's take a point on the ground just next to point of contact with ground and its velocity will also be zero and the trajectory of all other particles on disk from it at an instant will be same as that of point of contact of disk with ground. Then as the point on ground is inertial then point of contact of disk on ground should also be inertial.
 A: I'd say that the axis of rotation is a geometrical object, not a material one, which moves along with the point of contact of the disk and the floor.
It has nothing (or very little) to do with the points of the disk, except that the position of the axis is the projection of the center of the disk onto the floor.
A: 
My teacher told that instantaneous axis of rotation of a disk in pure rolling that is point of contact of disk with ground is an non inertial frame due to radial acceleration.

This doesn’t make sense. An axis is a single 1D line, but a reference frame is fully 4D. An axis cannot be a reference frame, inertial or non-inertial.
However, it is possible to construct a full 4D rotating reference frame in which the entire wheel is momentarily at rest. This reference frame will momentarily have the axis at the point of contact with the ground.
A: You can specify a non-inertial reference frame in which the disk is at rest; this frame is rotating with respect to an inertial frame. The origin of this reference frame could be selected to be at the center of mass (CM) or at a point on the edge of the disk.  In the non-inertial frame, fictitious forces need to be considered.
The typical approach is to consider the rotational motion in an inertial frame with respect to a specific point that is moving (in general accelerating) in the inertial frame.  Typically, that point is taken as the CM, but it can be another point such as one on the surface, moving and always in contact with the disk. For the point taken as the CM, the sum of the external torques about the CM is the change in angular momentum of the disk with respect to the CM.  For a point other than the CM, the relationship between torque and the change in angular momentum must also consider another term (the cross product of the velocity of the moving point in the inertial frame with the linear momentum of the CM).  An online reference, Dynamics by Dennis M. Kochmann provides a detailed discussion.
In general, the disk may be rolling down an incline and may slip. Pure rolling (no slip) is a special case.  In this case the point of contact of the disk with the surface is instantaneously at rest.
For pure rolling, the velocity of the moving point on the surface, always in contact with the rolling disk, in the inertial frame is zero.  We obtain the same result for the rotational motion using either this point or using the CM. This problem is solved in detail in the Kochmann reference, using various points of reference, all viewed from an inertial frame.
With respect to your question, the fact that the point on the surface in contact with the disk is instantaneously at rest makes it seem as if this is equivalent to an inertial system fixed with origin at the point, but it is not since the point is moving (in general accelerating down an incline) as the disk rolls.
Also, for pure rolling, friction does no work and the motion can be evaluated using simple conservation of energy considerations.
A: The axis of rotation is a geometry concept that explains the distribution of velocities among the particles of a moving body.

If you place the coordinate origin on the center of rotation (point in 2D, axis in 3D) then the velocity of any point located at $\vec{r} = \pmatrix{x \\ y}$ is
$$ \vec{v} = \vec{\omega} \times \vec{r} = \pmatrix{ -\omega \,x \\ \omega\,y } $$
The center of rotation is a point in space where the extended body has zero velocity (as seen by an inertial frame). The extended body is a concept of a virtual body of zero mass that extends throughout all space and rides along the actual body.
The point itself might move about from one time frame to another time frame, even in a discontinuous fashion, here at one instant and somewhere else at the next instant. It is not subject to "acceleration" as Newton's laws do not apply to it.
Consider another example. The motion of the helicopter blades during flight.

You can explain the motion by using a center of rotation offset to the side that is co-moving with the helicopter.

The rotation center although it is moving with the helicopter, it is not part of the helicopter and not subject to the radial acceleration the blade particles would have in this location.
Think of it as at each time frame there exists one point in the plane (or one axis in space) that you can designate as the instantenous axis of rotation and describe the motion of the object as a rotation about this point.
At each time frame, and from one time frame to another that point might be different completely.
