Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Fixed minor typos #2

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion 100-foundations.ipynb
Original file line number Diff line number Diff line change
Expand Up @@ -486,7 +486,7 @@
"# Model Selection and Regularization\n",
"\n",
"- we can choose many different kinds of models $f_\\theta$, $\\theta\\in\\mathbb{R}^d$\n",
" - small $d$: good predictions, but batch match to the data\n",
" - small $d$: good predictions, but bad match to the data\n",
" - large $d$: bad predictions, but good match to the data\n",
"- \"bias/variance tradeoff\", VC-dimension, empirical risk\n",
"- general approaches to dealing with this:\n",
Expand Down
2 changes: 1 addition & 1 deletion 120-convnet-ocr.ipynb
Original file line number Diff line number Diff line change
Expand Up @@ -118,7 +118,7 @@
"Structure:\n",
"\n",
"- perform 2D convolutions over the entire image\n",
"- assume that that has extracted features that correspond to characters\n",
"- assume that the network has extracted features that correspond to characters\n",
"- project those features into a 1D sequence\n",
"- classify the projected 1D feature sequence into characters\n",
"- perform training using EM alignment (CTC, Viterbi, etc.)"
Expand Down
2 changes: 1 addition & 1 deletion 210-pytorch.ipynb
Original file line number Diff line number Diff line change
Expand Up @@ -1453,7 +1453,7 @@
"# Future?\n",
"\n",
"- both PyTorch and Tensorflow have serious limitations\n",
"- either they will substantially, or new frameworks will come around"
"- either they will substantially change, or new frameworks will come around"
]
}
],
Expand Down
2 changes: 1 addition & 1 deletion 220-dataloader.ipynb
Original file line number Diff line number Diff line change
Expand Up @@ -1149,7 +1149,7 @@
"# Large Scale Distributed Training\n",
"\n",
"- commonly 16-256 GPUs training from the same dataset\n",
"- 200-1000 GB/s input data rate per GPU\n",
"- 200-1000 MB/s input data rate per GPU\n",
"- disk : GPU ratio of 2:1 to 10:1\n",
"- CPU node : GPU node ratio of 2:1\n",
"- on average, each disk only used by at most one GPU\n",
Expand Down
2 changes: 1 addition & 1 deletion 300-localization.ipynb
Original file line number Diff line number Diff line change
Expand Up @@ -85,7 +85,7 @@
"\n",
"Expectation-Maximization is a general approach to learning when some variable we need for recognition is missing from the training data.\n",
"\n",
"For OCR, the missing information is the _segmentation_ (e.g., character locations.\n",
"For OCR, the missing information is the _segmentation_ (e.g., character locations).\n",
"\n",
"EM approach:\n",
"- make a first guess at the segmentation\n",
Expand Down