Add models, yaml cards and tensorboard logs

data/JetClassII/JetClassII_ak4unscaled_full.yaml

@@ -0,0 +1,133 @@
+selection:
+   ### use `&`, `|`, `~` for logical operations on numpy arrays
+   ### can use functions from `math`, `np` (numpy), and `awkward` in the expression
+   (jet_pt > 15) & (jet_pt < 1000) & (np.abs(jet_eta) < 5) & (jet_label < 23)
+
+new_variables:
+   ### [format] name: formula
+   ### can use functions from `math`, `np` (numpy), and `awkward` in the expression
+   part_mask: ak.ones_like(part_energy)
+
+   # ## calculating scaled vectors
+   # part_px_scale: part_px / jet_pt * 500
+   # part_py_scale: part_py / jet_pt * 500
+   # part_pz_scale: part_pz / jet_pt * 500
+   # part_energy_scale: part_energy / jet_pt * 500
+
+   part_pt: np.hypot(part_px, part_py)
+   part_pt_log: np.log(part_pt)
+   part_e_log: np.log(part_energy)
+   #  part_pt_scale: np.hypot(part_px_scale, part_py_scale)
+   #  part_pt_scale_log: np.log(part_pt_scale)
+   #  part_e_scale_log: np.log(part_energy_scale)
+   part_logptrel: np.log(part_pt/jet_pt)
+   part_logerel: np.log(part_energy/jet_energy)
+   part_deltaR: np.hypot(part_deta, part_dphi)
+   part_d0: np.tanh(part_d0val)
+   part_dz: np.tanh(part_dzval)
+
+   # classes for reweighting
+   label_B: (((jet_label >= 0) & (jet_label < 2)) | (jet_label == 17))
+   label_C: (((jet_label >= 2) & (jet_label < 4)) | (jet_label == 18))
+   label_S: (((jet_label >= 4) & (jet_label < 6)) | (jet_label == 19))
+   label_D: (((jet_label >= 6) & (jet_label < 8)) | (jet_label == 20))
+   label_U: (((jet_label >= 8) & (jet_label < 10)) | (jet_label == 21))
+   label_G: ((jet_label == 10) | (jet_label == 22))
+   label_e: ((jet_label >= 11) & (jet_label < 13))
+   label_m: ((jet_label >= 13) & (jet_label < 15))
+   label_tauh: ((jet_label >= 15) & (jet_label < 17))
+
+preprocess:
+  ### method: [manual, auto] - whether to use manually specified parameters for variable standardization
+  method: manual
+  ### data_fraction: fraction of events to use when calculating the mean/scale for the standardization
+  data_fraction: 0.5
+
+inputs:
+   pf_points:
+      length: 128
+      pad_mode: wrap
+      vars: 
+         - [part_deta, null]
+         - [part_dphi, null]
+   pf_features:
+      length: 128
+      pad_mode: wrap
+      vars: 
+      ### [format 1]: var_name (no transformation)
+      ### [format 2]: [var_name, 
+      ###              subtract_by(optional, default=None, no transf. if preprocess.method=manual, auto transf. if preprocess.method=auto), 
+      ###              multiply_by(optional, default=1), 
+      ###              clip_min(optional, default=-5), 
+      ###              clip_max(optional, default=5), 
+      ###              pad_value(optional, default=0)]
+         - [part_pt_log, 1.7, 0.7]
+         - [part_e_log, 2.0, 0.7]
+         - [part_logptrel, -4.7, 0.7]
+         - [part_logerel, -4.7, 0.7]
+         - [part_deltaR, 0.2, 4.0]
+         - [part_charge, null]
+         - [part_isChargedHadron, null]
+         - [part_isNeutralHadron, null]
+         - [part_isPhoton, null]
+         - [part_isElectron, null]
+         - [part_isMuon, null]
+         - [part_d0, null]
+         - [part_d0err, 0, 1, 0, 1]
+         - [part_dz, null]
+         - [part_dzerr, 0, 1, 0, 1]
+         - [part_deta, null]
+         - [part_dphi, null]
+   pf_vectors:
+      length: 128
+      pad_mode: wrap
+      vars: 
+         - [part_px, null]
+         - [part_py, null]
+         - [part_pz, null]
+         - [part_energy, null]
+   pf_mask:
+      length: 128
+      pad_mode: constant
+      vars: 
+         - [part_mask, null]
+
+labels:
+   ### type can be `simple`, `custom`
+   ### [option 1] use `simple` for binary/multi-class classification, then `value` is a list of 0-1 labels
+   ### [option 2] otherwise use `custom` to define the label, then `value` is a map
+   type: custom
+   value: 
+      truth_label: jet_label 
+
+   ### notes on `jet_label`: 
+   ### - ranges from 0-22; 
+   ### - a full list of label names: [label_b, label_bbar, label_c, label_cbar, label_s, label_sbar, label_d, label_dbar, label_u, label_ubar, label_g, label_em, label_ep, label_mm, label_mp, label_tauhm, label_tauhp, label_bbbar, label_ccbar, label_ssbar, label_ddbar, label_uubar, label_gg]
+
+observers:
+   - jet_pt
+   - jet_eta
+   - jet_phi
+   - jet_energy
+   - jet_nparticles
+   - jet_mass
+
+weights:
+   ### [option 1] use precomputed weights stored in the input files
+   # use_precomputed_weights: true
+   # weight_branches: [weight, class_weight]
+   ### [option 2] compute weights on-the-fly using reweighting histograms
+   use_precomputed_weights: false
+   reweight_method: flat
+   reweight_vars:
+      jet_pt: [15, 20, 26, 35, 46, 61, 80, 106, 141, 186, 247, 326, 432, 571, 756, 1000]
+      jet_eta: [-10, 10]
+   reweight_classes:
+      [
+         label_B, label_C, label_S, label_D, label_U, label_G, label_e, label_m, label_tauh,
+      ]
+   class_weights:
+      [
+         1.0, 1.0, 1.0, 0.5, 0.5, 1.0, 0.5, 0.5, 1.0,
+      ]
+   reweight_hists:

models/JetClassII_SophonAK4/model.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:df06a3246907dcc8e19d1263eb0e36d9e06de997b7534baf176c7e3672325566
+size 2497916

models/JetClassII_SophonAK4/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8b32a2e71f1f3a3921f0886eda496ee85937896a8d654d20cc33783e016afe24
+size 2271841

tb_logs/events.out.tfevents.trainval.000

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cdbd125c9ab9db4bf221548b1ef987b2e880682d6f15f7991c4dc3c3b5502ab8
+size 3978173

tb_logs/events.out.tfevents.trainval.001

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e6f00f991a2d68a2044bfc9dfce6df12ad0ede55eea82f84110ed3fba6f406f6
+size 3978173