We present the best 265 sampled R-band light curves of spectroscopically identified Type Ia supernovae (SNe) from the Palomar Transient Factory (PTF; 2009-2012) survey and the intermediate Palomar Transient Factory (iPTF; 2013-2017). A model-independent light-curve template is built from our data-set with the purpose to investigate average properties and diversity in our sample. We searched for multiple populations in the light-curve properties using machine learning tools. We also utilized the long history of our light curves, up to 4000 days, to exclude any significant pre- or post- supernova flares. From the shapes of light curves we found the average rise time in the R band to be 16.8^+0.5^_-0.6_days. Although PTF/iPTF were single-band surveys, by modelling the residuals of the SNe in the Hubble-Lemaitre diagram, we estimate the average colour excess of our sample to be <E(B-V)>~0.05(2)mag and thus the mean corrected peak brightness to be M_R_=-19.02+/-0.02+5log(H_0_[km/s/Mpc]/70)mag with only weak dependence on light-curve shape. The intrinsic scatter is found to be {sigma}_R_=0.186+/-0.033mag for the redshift range 0.05<z<0.1, without colour corrections of individual SNe. Our analysis shows that Malmquist bias becomes very significant at z=0.13. A similar limitation is expected for the ongoing Zwicky Transient Facility (ZTF) survey using the same telescope, but new camera expressly designed for ZTF.