Disable rounding for float conversion

[rachitnigam]

Fixes #2336. Currently running into the same problem as #2335 (comment)

@ekiwi how did you fix it on the Rust side?

[rachitnigam]

I am currently getting this output:

{
  "inp": [
    15.123900413513184
  ],
  "out": [
    15.123900413513184,
    0.5600000023841858
  ]
}

Which might be fine since that is the closest representable number from the input files? Thoughts @ekiwi?

[ekiwi]

how did you fix it on the Rust side?

I fixed it by printing out the f32 instead of the f64

[ekiwi]

Stackoverflow seems to suggest using numpy to work with single precision float. However, it is still unclear to me whether numpy can actually print 32-bit floats.

https://stackoverflow.com/questions/2232362/correct-way-to-emulate-single-precision-floating-point-in-python

[ekiwi]

Awesome. Thanks!