mirror of
https://github.com/harvard-edge/cs249r_book.git
synced 2026-04-30 01:29:07 -05:00
0048bb6efa
RESTORED FILES: Configuration files: - .all-contributorsrc (contributor management) - .codespell-ignore (spelling check exclusions) - .lycheeignore (link checking exclusions) - .markdownlintignore (markdown linting exclusions) - .mdlintconfig.yml (markdown linting configuration) - .vale.ini (prose linting configuration) Documentation files: - BUILD.md (build instructions) - CODE_OF_CONDUCT.md (community guidelines) - DEVELOPMENT.md (development setup) - MAINTENANCE_GUIDE.md (maintenance procedures) - contribute.md (contribution guidelines) RATIONALE: - Preserves essential development tooling configuration - Ensures linting/quality checks remain functional across branches - Maintains documentation availability when switching back to dev branch - Balances clean project structure with development workflow needs
110 lines
975 B
Plaintext
110 lines
975 B
Plaintext
ehr
|
|
mor
|
|
appen
|
|
curren
|
|
relu
|
|
softmax
|
|
sigmoid
|
|
dropout
|
|
backprop
|
|
backpropagation
|
|
argmax
|
|
argmin
|
|
logits
|
|
dataset
|
|
subdataset
|
|
als
|
|
panting
|
|
sting
|
|
overfitting
|
|
underfitting
|
|
pretraining
|
|
finetuning
|
|
upweight
|
|
downweight
|
|
resnet
|
|
imagenet
|
|
cifar
|
|
tokenizer
|
|
tokenization
|
|
dequantize
|
|
quantized
|
|
quantization
|
|
dequantization
|
|
postprocessing
|
|
preprocessing
|
|
cnn
|
|
rnn
|
|
lstm
|
|
mlp
|
|
gpt
|
|
bert
|
|
encoder
|
|
decoder
|
|
layernorm
|
|
batchnorm
|
|
transformer
|
|
selfattention
|
|
feedforward
|
|
activation
|
|
positional
|
|
multihead
|
|
multilabel
|
|
multiclass
|
|
gpu
|
|
tpu
|
|
fpga
|
|
asic
|
|
bitwidth
|
|
compute
|
|
microbenchmark
|
|
throughput
|
|
latency
|
|
runtime
|
|
kernels
|
|
kernelized
|
|
parallelizable
|
|
interconnect
|
|
multichip
|
|
multinode
|
|
hardwareaware
|
|
pytorch
|
|
tensorflow
|
|
jax
|
|
onnx
|
|
xla
|
|
tvm
|
|
huggingface
|
|
mlir
|
|
nvidia
|
|
cuda
|
|
tensorcore
|
|
bfloat16
|
|
int8
|
|
float16
|
|
l2
|
|
l1
|
|
normed
|
|
regularizer
|
|
regularization
|
|
lossless
|
|
sparsifier
|
|
sparsification
|
|
quantifier
|
|
approximate
|
|
approximation
|
|
binarization
|
|
ternarization
|
|
convergence
|
|
convexity
|
|
nondeterministic
|
|
markdown
|
|
qmd
|
|
tikz
|
|
quarto
|
|
yaml
|
|
precommit
|
|
boxS
|
|
Handlin
|
|
Dota
|
|
ALine |