stvhuang
stvhuang
+1, I met the same lower part error. on both macOS 10.14 and ArchLinux with Ruby 2.7.0. Edit: Solved by downgrading `pdfkit` to 0.8.4.
@SeaYM Try downgrade your PyTorch version to `1.6.0`.
Latest version of Pyright (`1.1.262`) starts to throw the similar errors. (`"batch_size" is not defined`)
I got this exception when running out of gpu memory.
> 实际上线性无关的定义应该是 **不存在** 一组标量使得一组向量形成某种线性关系（或者说这组标量可以 **使得这组向量之和为0**），则说明这一组向量是线性无关的。 AFAIK, the definition of linear independence in the book is correct. See [wiki's definition.](https://en.wikipedia.org/wiki/Linear_independence#Definition) And actually, your argument has a flaw, with $\Lambda=[0,0,\dots,0]$, $\lambda_1\mathbf{v}_1+\lambda_2\mathbf{v}_2+\dots+\lambda_n\mathbf{v}_n=\mathbf{0}$ will...
I think the vector concatenation here refers to **[x_t \\\\ h_{t-1}]** in Equation 6.57. Also, the **W** in Equation 6.57 is a bit ambiguous. It is **[W U]** for **W**...
Hey, @ningpingyang, do you mean that **z in (a)** and **x in (b)** should share the same size, and also **x in (a)** and **z in (b)** should share the...
I think "t" here refers to the t-th weights update instead of the t-th epoch; therefore, it is not an error.
I am convinced. I didn't notice there is a 'k.' 😅