Repository navigation

Refactor - #2

Closed

magana272 wants to merge 186 commits into

mainfrom

Commits on Apr 14, 2026

memorymapping
magana272
committed
4073022
n_trials smaller
magana272
committed
6b7a427
batchsize: smaller
magana272
committed
51e05cf
batchsize: smaller
magana272
committed
af8d5b2
batchsize: smaller
magana272
committed
02fa957
loader: dont move entire dataset to gpu
magana272
committed
e7b7b3a
optimizer
magana272
committed
b04d81f
optimizer
magana272
committed
b8e2690
optimizer
magana272
committed
4a4bf82
thread, and faster dataloader
magana272
committed
0024fe9
thread: or not to thread
magana272
committed
32dff48
thread: remove on copy
magana272
committed
4446c9b
numworkers reduced
magana272
committed
b1b250a
fix: load
magana272
committed
770b2f7
fix: train builder
magana272
committed
3d6bc39
fix: train builder
magana272
committed
e681d66
fix: train builder
magana272
committed
d7d8d77
RUN IN GOOGLE COLAB
magana272
committed
ef53e49
update batch and epock
magana272
committed
5228e61
Merge branch 'experiment' of https://github.com/magana272/CellTypeClassification into experiment
magana272
committed
19a87cb
Requirements
magana272
committed
85f0508

Commits on Apr 15, 2026

fix
magana272
committed
b79424b
preload all on GPU
magana272
committed
0364ee1
preload all on GPU
magana272
committed
7809eff
preload all on GPU
magana272
committed
5090f33
increase batch size
magana272
committed
11694ef
back to cpu bound
magana272
committed
6df9b62
remove double to
magana272
committed
8c8795b
simplify CNN
magana272
committed
ec8a717
estimate size
magana272
committed
b35c154
samll batch
magana272
committed
fb6be66
one trail
magana272
committed
08cbd53
SGD
magana272
committed
d8bdd9a
larger batch size
magana272
committed
921abbb
larger batch and more trials
magana272
committed
c6eb4db
shrick model
magana272
committed
4d6741d
increase model size
magana272
committed
77c971f
integratepreprocessing
magana272
committed
2d8efeb
transformer gets all genes
magana272
committed
180381d
update batch size and hps
magana272
committed
1cf08b4
cell search hyperparameter
magana272
committed
4f0cb23
cell search hyperparameter
magana272
committed
a86eec8
hyperparamter tuning and cross dataset eval
magana272
committed
0fcb169
update epoch
magana272
committed
b90ba44
modelstate change
magana272
committed
1ec1271
Merge branch 'experiment' of https://github.com/magana272/Cell-Type-Classification into experiment
magana272
committed
b121499
Run_the_models
magana272
committed
718190c
cuda failing in some env
magana272
committed
03217fc
Merge branch 'experiment' of https://github.com/magana272/Cell-Type-Classification into experiment
magana272
committed
17cb3dc
log epoch and fix cuda error
magana272
committed
06c8af6
remove prgress bar
magana272
committed
09e4d5b
faiss: can't see to get to work with cuda13
magana272
committed
80b741f
knn work around
magana272
committed
762ae10
knn work around
magana272
committed
a2f81c4
fix defualting
magana272
committed
2b44cde
torch knn
magana272
committed
4106bf8
torch knn
magana272
committed
2aec8a7
tensor knn
magana272
committed
526167c
tensor knn
magana272
committed
b14423e
bottleneck
magana272
committed
eaf41b9
hidden dim
magana272
committed
eab19be
hps graph
magana272
committed
7f2c799
more trials
magana272
committed
a4314a7

Commits on Apr 16, 2026

gnn model state fix
magana272
committed
f4596d4
saveing results, comparison, and interpert ability
magana272
committed
e811adc
visualize of 10x dataset
magana272
committed
c393895
larger batch size
magana272
committed
1a828b0
visulization updates, rich requirement, gpu preprocessing
magana272
committed
00599ef
git normalization is a hyper parameter, also remove print.
magana272
committed
9878aaf
must be float
magana272
committed
190d652
Final workflow
magana272
committed
18f7f90
increase trails
magana272
committed
4221af1
Merge branch 'experiment' of https://github.com/magana272/Cell-Type-Classification into experiment
magana272
committed
0c08bd3
increase transformer batch
magana272
committed
c9735cd
increase epoch and trail
magana272
committed
a5f9a61
lower the patience
magana272
committed
1879eca
smaller epoch and lower the patience
magana272
committed
899e373
move to gpu and fix epoch size
magana272
committed
6523bd4
hyper search fix
magana272
committed
8cd055d
batch size
magana272
committed
12542f1
fix batch
magana272
committed
caa477b
smaller search
magana272
committed
7ea4e77
opt tran
magana272
committed
cee947d
narrow search space
magana272
committed
2846b85
narrow search space
magana272
committed
7cffcff
tune epoch
magana272
committed
9afb1ed
fix transformer hyper paramter
magana272
committed
e87f54c
increase batch
magana272
committed
327212e
decrease trial
magana272
committed
4180d30
increase batch
magana272
committed
532b6f8
trans
magana272
committed
d2ce2f8
norm error
magana272
committed
adfb2e5
workers
magana272
committed
9b098da
figures
magana272
committed
438ec93
threads
magana272
committed
ecfa9b4
mlp fix
magana272
committed
4de85f0
results
magana272
committed
a62e38c
unkown cell
magana272
committed
aeaa9db
unkown cell
magana272
committed
f3c1cb6
unkown pipeline
magana272
committed
bcbbfb2
umap embedding result
magana272
committed
b614c48
Created using Colab
magana272
committed
58e254b
no norm
magana272
committed
30f8cfd
Merge branch 'experiment' of https://github.com/magana272/Cell-Type-Classification into experiment
magana272
committed
1d090ec
experiment
magana272
committed
3912fbe
test test
magana272
committed
4d8a388
tos
magana272
committed
4c881db

Commits on Apr 18, 2026

Created using Colab
magana272
committed
02d58d2
Default model comparison
magana272
committed
1df25dd
Merge branch 'experiment' of https://github.com/magana272/Cell-Type-Classification into experiment
magana272
committed
0274dc9
import error
magana272
committed
757549d
imports
magana272
committed
58dbe6e
im
magana272
committed
9b3a54f
import
magana272
committed
5cb9b41
TOSICA: models integrated
magana272
committed
0845e4c
TOSICA: models integrated
magana272
committed
c446cd4
ana data
magana272
committed
1822de0
dense
magana272
committed
02e41cd
adata:load and add meta data
magana272
committed
d56fc65
h5ad file
magana272
committed
ca142c4
print
magana272
committed
f080aec
split kills ram
magana272
committed
700077e
split
magana272
committed
4e1a44b
oom err
magana272
committed
0aeeeec
oom: error
magana272
committed
5f55af5
sample
magana272
committed
f3fe5fe
path
magana272
committed
f0f2431
path
magana272
committed
a6c848c
gene
magana272
committed
ca338c8
batch
magana272
committed
77a1484
batch size
magana272
committed
809a338
batch size
magana272
committed
2905acc
batch size
magana272
committed
e8e54e6
batch size
magana272
committed
fcf41ba
batch size
magana272
committed
e6b93f4
batch size
magana272
committed
2c8dad4
batch size
magana272
committed
41e691c
batch size
magana272
committed
58fe188
batch size
magana272
committed
0e9aaf0
originalmodel
magana272
committed
6e00b8f
cuda error
magana272
committed
078b7c0
Merge branch 'experiment' of https://github.com/magana272/Cell-Type-Classification into experiment
magana272
committed
b167002
model comp
magana272
committed
bf7f683
model comp
magana272
committed
82174a2
refactor
magana272
committed
bdb0a19
label
magana272
committed
e0ebd47
refactored
magana272
committed
de385c4
done
magana272
committed
3cd6a62
done
magana272
committed
ba2fbf1

Commits on Apr 19, 2026

refactor
magana272
committed
8f466c8
validate
magana272
committed
3141cda
data
magana272
committed
f57d584
datasets
magana272
committed
f146024
update
magana272
committed
b2533db
fixed link
magana272
committed
1d9de6a
mbrain
magana272
committed
0c5dc97
mbrain
magana272
committed
895def3
fixed
magana272
committed
a3afaed
brain
magana272
committed
766956d
brain
magana272
committed
8745af0
brain
magana272
committed
431c820
pickle error
magana272
committed
b86f331
download
magana272
committed
c66f77c
sparse
magana272
committed
4f71790
sparse error
magana272
committed
49f1162
sparse error
magana272
committed
baf9029
m
magana272
committed
328506a
test
magana272
committed
f23dfbf
nvg
magana272
committed
6bec188
HVG
magana272
committed
8f5355c
equal learning rates
magana272
committed
de69fad

Commits on Apr 20, 2026

error fix
magana272
committed
05b7df0
error import
magana272
committed
7824e8c
fixed
magana272
committed
6116428
atten error
magana272
committed
8d2d108
balanced population
magana272
committed
5fa4049
hvg
magana272
committed
709fd76
batch size
magana272
committed
2d5b2f7
don't need pos encoding
magana272
committed
bdf901d
reactome
magana272
committed
5b49f87
mouse
magana272
committed
4af2f37
save results
magana272
committed
ac795a7
run all
magana272
committed
e3ad2e9
reorder script
magana272
committed
775708f
increase batch size
magana272
committed
815deca

Commits on Apr 21, 2026

Created using Colab
magana272
committed
8da9c6b