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.

Sign up for GitHub

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

Jump to bottom

Added higgs boson tabnet examples #1209

Merged
merged 1 commit into from
Jun 15, 2021
Merged

Added higgs boson tabnet examples #1209

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
87 changes: 87 additions & 0 deletions examples/tabnet/higgs/medium_config.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,87 @@
input_features:
- name: lepton_pT
type: numerical
- name: lepton_eta
type: numerical
- name: lepton_phi
type: numerical
- name: missing_energy_magnitude
type: numerical
- name: missing_energy_phi
type: numerical
- name: jet_1_pt
type: numerical
- name: jet_1_eta
type: numerical
- name: jet_1_phi
type: numerical
- name: jet_1_b-tag
type: numerical
- name: jet_2_pt
type: numerical
- name: jet_2_eta
type: numerical
- name: jet_2_phi
type: numerical
- name: jet_2_b-tag
type: numerical
- name: jet_3_pt
type: numerical
- name: jet_3_eta
type: numerical
- name: jet_3_phi
type: numerical
- name: jet_3_b-tag
type: numerical
- name: jet_4_pt
type: numerical
- name: jet_4_eta
type: numerical
- name: jet_4_phi
type: numerical
- name: jet_4_b-tag
type: numerical
- name: m_jj
type: numerical
- name: m_jjj
type: numerical
- name: m_lv
type: numerical
- name: m_jlv
type: numerical
- name: m_bb
type: numerical
- name: m_wbb
type: numerical
- name: m_wwbb
type: numerical
output_features:
- name: label
type: binary
weight_regularization: null
combiner:
type: tabnet
size: 32 # N_a
output_size: 96 # N_d
sparsity: 0.000001 # lambda_sparse
bn_virtual_divider: 32 # factor to divide batch_size B to get B_v from the paper
bn_momentum: 0.9 # m_B
num_steps: 8 # N_steps
relaxation_factor: 2 # gamma
use_keras_batch_norm: false
bn_virtual_bs: 256 # B_v
training:
batch_size: 8192 # B
eval_batch_size: 500000 # 65536 131072 262144 524288
epochs: 1000
early_stop: 20
learning_rate: 0.025
optimizer:
type: adam
decay: true
decay_steps: 10000
decay_rate: 0.9
staircase: true
regularization_lambda: 1
validation_field: label
shuffle_buffer_size: 1500000
86 changes: 86 additions & 0 deletions examples/tabnet/higgs/small_config.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,86 @@
input_features:
- name: lepton_pT
type: numerical
- name: lepton_eta
type: numerical
- name: lepton_phi
type: numerical
- name: missing_energy_magnitude
type: numerical
- name: missing_energy_phi
type: numerical
- name: jet_1_pt
type: numerical
- name: jet_1_eta
type: numerical
- name: jet_1_phi
type: numerical
- name: jet_1_b-tag
type: numerical
- name: jet_2_pt
type: numerical
- name: jet_2_eta
type: numerical
- name: jet_2_phi
type: numerical
- name: jet_2_b-tag
type: numerical
- name: jet_3_pt
type: numerical
- name: jet_3_eta
type: numerical
- name: jet_3_phi
type: numerical
- name: jet_3_b-tag
type: numerical
- name: jet_4_pt
type: numerical
- name: jet_4_eta
type: numerical
- name: jet_4_phi
type: numerical
- name: jet_4_b-tag
type: numerical
- name: m_jj
type: numerical
- name: m_jjj
type: numerical
- name: m_lv
type: numerical
- name: m_jlv
type: numerical
- name: m_bb
type: numerical
- name: m_wbb
type: numerical
- name: m_wwbb
type: numerical
output_features:
- name: label
type: binary
weight_regularization: null
combiner:
type: tabnet
size: 24 # N_a
output_size: 26 # N_d
sparsity: 0.000001 # lambda_sparse
bn_virtual_divider: 32 # factor to divide batch_size B to get B_v from the paper
bn_momentum: 0.6 # m_B
num_steps: 5 # N_steps
relaxation_factor: 1.5 # gamma
use_keras_batch_norm: false
bn_virtual_bs: 512 # B_v
training:
batch_size: 16384 # B
eval_batch_size: 500000 # 65536 131072 262144 524288
epochs: 1000
early_stop: 20
learning_rate: 0.02
optimizer:
type: adam
decay: true
decay_steps: 20000
decay_rate: 0.9
staircase: true
regularization_lambda: 1
validation_field: label
16 changes: 16 additions & 0 deletions examples/tabnet/higgs/train_higgs_medium.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,16 @@
import logging

from ludwig.api import LudwigModel
from ludwig.datasets import higgs

model = LudwigModel(
config='medium_config.yaml',
logging_level=logging.INFO,
)

higgs_df = higgs.load()
model.train(
dataset=higgs_df,
experiment_name='higgs_medium',
model_name='higgs_tabnet_medium'
)
16 changes: 16 additions & 0 deletions examples/tabnet/higgs/train_higgs_small.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,16 @@
import logging

from ludwig.api import LudwigModel
from ludwig.datasets import higgs

model = LudwigModel(
config='small_config.yaml',
logging_level=logging.INFO,
)

higgs_df = higgs.load()
model.train(
dataset=higgs_df,
experiment_name='higgs_small',
model_name='higgs_tabnet_small'
)