update fip

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "causica"
-version = "0.4.3"
+version = "0.4.4"
 description = ""
 readme = "README.md"
 authors = ["Microsoft Research - Causica"]
@@ -64,13 +64,15 @@ known_first_party = ["causica"]
 treat_comments_as_code = ["# %%"]
 
 [tool.pytest.ini_options]
+norecursedirs = "integration"
 addopts = "--durations=200"
 junit_family = "xunit1"
 
 [tool.mypy]
 ignore_missing_imports = true
 exclude = [
-  "research_experiments/avid"
+  "research_experiments/avid",
+  "research_experiments/fip",
 ]
 
 [tool.pylint.main]

research_experiments/fip/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "fip"
-version = "0.1.0"
+version = "0.1.1"
 description = "A Fixed-Point Approach for Causal Generative Modeling"
 readme = "README.md"
 authors = ["Meyer Scetbon", "Joel Jennings", "Agrin Hilmkil", "Cheng Zhang", "Chao Ma"]
@@ -11,8 +11,34 @@ license = "MIT"
 
 [tool.poetry.dependencies]
 python = "~3.10"
-causica = "0.4.1"
+causica = "0.4.2"
+
+[tool.poetry.group.dev.dependencies]
+black = {version="^22.6.0", extras=["jupyter"]}
+isort =  "^5.10.1"
+jupyter = "^1.0.0"
+jupytext = "^1.13.8"
+mypy = "^1.0.0"
+pre-commit = "^2.19.0"
+pylint = "^2.14.4"
+pytest = "^7.1.2"
+pytest-cov = "^3.0.0"
+seaborn = "^0.12.2"
+types-python-dateutil = "^2.8.18"
+types-requests = "^2.31.0.10"
+
 
 [build-system]
 requires = ["poetry-core>=1.0.0"]
 build-backend = "poetry.core.masonry.api"
+
+[tool.black]
+line-length = 120
+
+[tool.isort]
+line_length = 120
+profile = "black"
+py_version = 310
+known_first_party = ["fip"]
+# Keep import sorts by code jupytext percent block (https://github.com/PyCQA/isort/issues/1338)
+treat_comments_as_code = ["# %%"]

research_experiments/fip/src/fip/config/amortization/leaf_prediction.yaml

@@ -1,4 +1,5 @@
-seed_everything: 234  
+seed_everything: 234
+
 model:
   class_path: fip.tasks.amortization.leaf_prediction.LeafPrediction
   init_args:
@@ -14,6 +15,7 @@ model:
     num_to_keep_training: 10
     distributed: false 
     elimination_type: "self"
+
 trainer:
   max_epochs: 2000
   accelerator: gpu
@@ -22,16 +24,19 @@ trainer:
   profiler: "simple"
   devices: 1
   accumulate_grad_batches: 1
+
 best_checkpoint_callback:
   dirpath: "./outputs/"
   filename: "best_model"
   save_top_k: 1
   mode: "min"
   monitor: "val_loss"
   every_n_epochs: 1
+
 last_checkpoint_callback:
   save_last: true
   save_top_k: 0  # only the last checkpoint is saved
+
 early_stopping_callback:
   monitor: "val_loss"
   min_delta: 0.0001

research_experiments/fip/src/fip/config/scm_learning_with_ground_truth/scm_learning_true_graph.yaml

@@ -1,4 +1,5 @@
 seed_everything: 234
+
 model:
   class_path: fip.tasks.scm_learning_with_ground_truth.scm_learning_true_graph.SCMLearningTrueGraph
   init_args:
@@ -9,13 +10,14 @@ model:
     num_heads: 8
     d_feedforward: 128
     total_nodes: 4
-    total_layers: 2
+    total_layers: 1
     dropout_prob: 0.
     mask_type: "none"
     attn_type: "causal"
     cost_type: "dot_product"
     learnable_loss: false
     distributed: false
+
 trainer:
   max_epochs: 1000
   accelerator: gpu
@@ -24,19 +26,22 @@ trainer:
   log_every_n_steps: 10
   inference_mode: false
   profiler: "simple"
+
 early_stopping_callback:
   monitor: "val_loss"
   min_delta: 0.0001
   patience: 500
   verbose: False
   mode: "min"
+
 best_checkpoint_callback:
   dirpath: "./outputs/"
   filename: "best_model"
   save_top_k: 1
   mode: "min"
   monitor: "val_loss"
   every_n_epochs: 1
+
 last_checkpoint_callback:
   save_last: true
   save_top_k: 0  # only the last checkpoint is saved

research_experiments/fip/src/fip/config/scm_learning_with_ground_truth/scm_learning_true_perm.yaml

@@ -1,4 +1,5 @@
 seed_everything: 5000
+
 model:
   class_path: fip.tasks.scm_learning_with_ground_truth.scm_learning_true_perm.SCMLearningTruePerm
   init_args:
@@ -16,6 +17,7 @@ model:
     cost_type: "dot_product"
     learnable_loss: false
     distributed: false
+
 trainer:
   max_epochs: 1000
   accelerator: gpu
@@ -24,19 +26,22 @@ trainer:
   log_every_n_steps: 10
   inference_mode: false
   profiler: "simple"
+
 early_stopping_callback:
   monitor: "val_loss"
   min_delta: 0.0001
   patience: 500
   verbose: False
   mode: "min"
+
 best_checkpoint_callback:
   dirpath: "./outputs/"
   filename: "best_model"
   save_top_k: 1
   mode: "min"
   monitor: "val_loss"
   every_n_epochs: 1
+
 last_checkpoint_callback:
   save_last: true
   save_top_k: 0  # only the last checkpoint is saved

research_experiments/fip/src/fip/config/scm_learning_with_predicted_truth/scm_learning_predicted_leaf.yaml

@@ -1,4 +1,5 @@
-seed_everything: 5000 
+seed_everything: 234 
+
 model:
   class_path: fip.tasks.scm_learning_with_predicted_truth.scm_learning_predicted_leaf.SCMLearningPredLeaf
   init_args:
@@ -18,26 +19,30 @@ model:
     cost_type: "dot_product"
     learnable_loss: false
     distributed: false
+
 trainer:
   max_epochs: 1000
   accelerator: gpu
   devices: 1
   check_val_every_n_epoch: 1
   log_every_n_steps: 10
   inference_mode: false
+
 early_stopping_callback:
   monitor: "val_loss"
   min_delta: 0.0001
   patience: 500
   verbose: False
   mode: "min"
+
 best_checkpoint_callback:
   dirpath: "./outputs/"
   filename: "best_model"
   save_top_k: 1
   mode: "min"
   monitor: "val_loss"
   every_n_epochs: 1
+
 last_checkpoint_callback:
   save_last: true
   save_top_k: 0  # only the last checkpoint is saved