george-colin
george-colin
I get the same 'nonsense' with the 7B model. I'm using ubuntu and an AMD cpu. The output looks so 'wrong' that I imagine we have got something completely messed...
gjmulder - could you give your md5sum values for the weights you downloaded please - e.g. consolidated.00.pth - then I can see if I am starting off with the right...
OK, now I get sensible results with 7B model. I just got the latest version from github and re-ran everything - previously my version was a couple of hours old....