It was after an Ar ion milling process right? If the milling process is long, it seems that the resist can overheat and then it gets harder to remove. I'm not sure if it's only overheating or maybe the hardening is due to some kind of ion implantation, or maybe implantation of re-scattered atoms from the material you were milling. You can perform an oxygen plasma etching to remove since TGMR is organic, as the previous response suggest. But in my experience, the resist in the border of the pattern remains after plasma etching. The cleanest result for me is using N-Methyl-2-Pyrrolidone, the usual chemical etching, for long time (several hours) and it's better to heat to 80°C too. But if this is not enough, I use ultrasonic wave cleaning (45 Hz) during several minutes. But you have to be sure that your pattern will not be destroyed...
If the resist still remains, you can try plasma etching but, again, the resist in the borders could remain. A lower power ultrasonic cleaning (maybe 100 Hz) could work after plasma etching to remove the resist remaining in the borders. Be careful, sometimes it can not be seen with the microscope but the resist in the borders is there