Did you notice that OPA657 is only stable with close loop gains >= 7? This means, that there's also an upper stability limit for the feedback capacitor, depending on the photodiode capacitance. 1pF seems somewhat high. The fact, that you observed oscillations at the supply pins seems to indicate insuitable placement of the bypassing capacitors. Personally, I would rather use 100 nF and high capacitance ceramic or tantal as suggested in the OPA657 datasheet. 1nF is possibly useful, the pF values are of little use. As the datasheet tells, 0.25" should be a maximum distance of 100 nF to supply pins, a solid ground plane is required.