Example performance -- is this input out of distribution?

[dillon-0]

Hello! Thanks for creating SMT and sharing the repo.

I tried running with the PRAIG/smt-fp-grandstaff and got results that were not very accurate. Here's the image and output I got. Notably the time signature wrong, is missing most measures, and can't render in https://verovio.humdrum.org/. I tried prepending **kern **kern to the output and pasting there but it says "Rhythm off by 1/8 at field 2.

Question: is this kind of image out of distribution / too low res? Or would you expect better performance for this kind of input?

Thanks!

*clefF4	*clefG2
*k[b-]	*k[b-]
*M2/4	*M2/4
8D	4cc
8F	16ddL[
.	16eeJ]
.	16ffJq
8E	2ee
=	=
4F	8eeLq
.	16eeJq
.	16ddJq
8FL	2cc
8G	.
8F	.
8FJ	.
=	=
4G	8ddL[
.	16eeq
.	16ddJ
8A	16ccL
.	16dd
.	16ccJ
4C	16cc
.	16ccJ
=	=
4F	2cc
8CL	.
8D	.
8E	4r
8DJ	.
=	=
4F	8aL
.	16cc
4r	16a
.	16cc
4GG	16b
.	16ccJ
=	=
4GG	16aL
.	16cc
4GG	16b
.	16dd
4EE	16b
.	16ddJ
=	=
*-	*-

[dillon-0]

Here's another example I tried. For this one, I'm wondering if the image is too distorted (not flat enough), or if you would expect good performance here

Would appreciate any thoughts you have @antoniorv6 ! Thanks