the formula of comp_beta in paper is inconsistent with code formula

[bobo-0731]

Thank you for your excellent work,and Im interested in your contribution，but now I have a doubt while reading the code now. The calculation of comp_beta β in the paper is as follows： β(r; σ, α) = ∑_k {Ti(σt)(1-exp(-σtiδi))αi} but models.rendering.py in line 95 is expressed comp_beta = (trans_weights[..., None] * beta).sum(dim=-2) Their representations are inconsistent，whether to replace beta with alpha in the code?