Skip to content

issue about "gpt.py" in code part of ch 4 #50

@976213951

Description

@976213951

In the line 66, it is supposed to be "nn.Linear(d_out, d_in)" instead of "nn.Linear(d_out, d_out)" since the linear attempts to project the dimension of the attention (d_out) back to the dimension of embedding "d_in". However, this code can work without any error report because the "d_out" is set equal to "d_in" by default.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions