mohamad maheri
|
030e40e851
|
tuning for clustering by lstm and data_batching_new (make clustering module depart from prediction network)
|
2 years ago |
mohamad maheri
|
c701e2c338
|
only consider positive interaction for task embedding
|
2 years ago |
mohamad maheri
|
5273533fc8
|
Revert "deep kmeans loss"
This reverts commit b4d130e2
|
2 years ago |
mohamad maheri
|
3a3b256e65
|
Revert "prepare for hyper parameter tuning"
This reverts commit 576c0e97
|
2 years ago |
mohamad maheri
|
b4d130e28e
|
deep kmeans loss
|
2 years ago |
mohamad maheri
|
acacb68d33
|
tanp loss (kl loss)
|
2 years ago |
mohamad maheri
|
d425ba0c97
|
cluster loss (try to hard assignment)
|
2 years ago |
mohamad maheri
|
576c0e97f2
|
prepare for hyper parameter tuning
|
2 years ago |
mohamad maheri
|
d2d5563db1
|
prepare for hyper parameter tuning
|
2 years ago |
mohamad maheri
|
dda2618d06
|
solve gamma and beta problem
|
2 years ago |
mohamad maheri
|
908149d81b
|
task encoder consider item's score
there is problems that should set inner loop larger than one
|
2 years ago |
mohamad maheri
|
d9e027a938
|
user clustering effects prediction network's weights
|
2 years ago |