Inconsistent Young’s Modulus (E) Scale in Histogram Figure and VLM Annotation Instructions (especially for metal)

[ltjldc]

Hi Pixie team,

I noticed two issues related to the Young’s modulus (E) values used in the project, and I would like to confirm whether they are intended or if they need correction.

---

1. Histogram figure shows E without powers of 10

In the attached screenshot from the paper, the x-axis values are around 10^3 to 10^8.

However, real metal Young’s modulus values are usually on the order of 10¹¹ Pa (for aluminum, roughly 70 GPa).
So I was confused because the figure does not show any peaks near 10¹⁰–10¹¹.

Question:
Is the histogram normalized, scaled, or downsampled for visualization?
Or is the dataset using reduced E values on purpose (for numerical stability in MPM)?

This part is unclear and I hope to confirm the intended interpretation.

---

2. Inconsistent E values inside vlm_seg_class_instruction.py

Inside the VLM material annotation code, the Young’s modulus ranges do not follow a consistent scale:

Soda cans

"E": [5e10, 8e10]

This matches typical aluminum stiffness (50–80 GPa).

Metal crates

"E": [8e7, 1.2e8]

This is 800–1200 times smaller than the soda can values, yet both objects are described as rigid metal.

Also, in the tips section for soda cans, it says:

Keep E relatively high (around 1e8) for metal stiffness

This directly contradicts the example above (5e10–8e10).

Questions:

1. Should all metal objects use the same order of magnitude (e.g., 1e10–1e11), and the crate values are accidentally too low?

2. Or is the lower E range intentional for simulation stability when the object is large?

3. Should the documentation / tips be updated to use a consistent scale?

---

I would appreciate clarification on:

• whether the histogram figure uses scaled E values,

• and whether the annotation code needs unification or if the mixed scales are intentional for MPM stability.

Thanks a lot for your time, and I really enjoy using Pixie.