While working through these very helpful documents right now, I noticed something that didn't make sense to me ---
When calculating the partial for the average w.r.t. out, shouldn't the first 10 in d_avg_d_out <- torch_tensor(10)$`repeat\`(10)$unsqueeze(1)$t() be replaced with 1/10?
The LaTeX for the chain rule also appears broken at the moment.
Thank you all for your hard work! Loving how powerful this framework is!
Document: https://torch.mlverse.org/technical/modules/
While working through these very helpful documents right now, I noticed something that didn't make sense to me --- When calculating the partial for the average w.r.t.
out
, shouldn't the first10
ind_avg_d_out <- torch_tensor(10)$`repeat\`(10)$unsqueeze(1)$t()
be replaced with1/10
?The LaTeX for the chain rule also appears broken at the moment.
Thank you all for your hard work! Loving how powerful this framework is!