update test

tests/e2e/test_e2e_skip_trained.py

@@ -1,4 +1,4 @@
-"""Emmental e2e with skipping learned."""
+"""Emmental e2e with skipping trained data."""
 import logging
 import shutil
 from functools import partial
@@ -27,7 +27,7 @@ def test_e2e_skip_trained(caplog):
     caplog.set_level(logging.INFO)
 
     dirpath = "temp_test_e2e_skip_trained"
-    use_exact_log_path = False
+    use_exact_log_path = True
     Meta.reset()
     init(dirpath, use_exact_log_path=use_exact_log_path)
 
@@ -149,24 +149,24 @@ def forward(self, input):
     ]
     # Build model
 
-    mtl_model = EmmentalModel(name="all", tasks=tasks)
+    model = EmmentalModel(name="all", tasks=tasks)
 
     # Create learner
     emmental_learner = EmmentalLearner()
 
     config = {
-        "meta_config": {"seed": 0, "verbose": False},
+        "meta_config": {"seed": 0, "verbose": True},
         "learner_config": {
-            "n_steps": 200,
+            "n_steps": 10,
             "epochs_learned": 0,
-            "steps_learned": 130,
-            "skip_learned_data": True,
+            "steps_learned": 0,
+            "skip_learned_data": False,
             "online_eval": True,
             "optimizer_config": {"lr": 0.01, "grad_clip": 100},
         },
         "logging_config": {
-            "counter_unit": "epoch",
-            "evaluation_freq": 0.2,
+            "counter_unit": "batch",
+            "evaluation_freq": 5,
             "writer_config": {"writer": "json", "verbose": True},
             "checkpointing": True,
             "checkpointer_config": {
@@ -177,39 +177,48 @@ def forward(self, input):
                 "checkpoint_runway": 1,
                 "checkpoint_all": False,
                 "clear_intermediate_checkpoints": True,
-                "clear_all_checkpoints": True,
+                "clear_all_checkpoints": False,
             },
         },
     }
     Meta.update_config(config)
 
     # Learning
     emmental_learner.learn(
-        mtl_model,
+        model,
         [train_dataloader, dev_dataloader],
     )
 
-    test_score = mtl_model.score(test_dataloader)
+    test_score = model.score(test_dataloader)
 
-    assert test_score["task1/synthetic/test/loss"] > 0.1
+    assert test_score["task1/synthetic/test/loss"] > 0.4
 
     Meta.reset()
     init(dirpath, use_exact_log_path=use_exact_log_path)
 
     config = {
-        "meta_config": {"seed": 0, "verbose": False},
+        "meta_config": {"seed": 0, "verbose": True},
         "learner_config": {
-            "n_steps": 200,
+            "n_steps": 40,
             "epochs_learned": 0,
-            "steps_learned": 0,
-            "skip_learned_data": False,
+            "steps_learned": 10,
+            "skip_learned_data": True,
             "online_eval": True,
             "optimizer_config": {"lr": 0.01, "grad_clip": 100},
+            "optimizer_path": (
+                f"{dirpath}/" "best_model_model_all_train_loss.optimizer.pth"
+            ),
+            "scheduler_path": (
+                f"{dirpath}/" "best_model_model_all_train_loss.scheduler.pth"
+            ),
+        },
+        "model_config": {
+            "model_path": f"{dirpath}/best_model_model_all_train_loss.model.pth"
         },
         "logging_config": {
-            "counter_unit": "epoch",
-            "evaluation_freq": 0.2,
-            "writer_config": {"writer": "tensorboard", "verbose": True},
+            "counter_unit": "batch",
+            "evaluation_freq": 5,
+            "writer_config": {"writer": "json", "verbose": True},
             "checkpointing": True,
             "checkpointer_config": {
                 "checkpoint_path": None,
@@ -219,20 +228,23 @@ def forward(self, input):
                 "checkpoint_runway": 1,
                 "checkpoint_all": False,
                 "clear_intermediate_checkpoints": True,
-                "clear_all_checkpoints": True,
+                "clear_all_checkpoints": False,
             },
         },
     }
     Meta.update_config(config)
 
+    if Meta.config["model_config"]["model_path"]:
+        model.load(Meta.config["model_config"]["model_path"])
+
     # Learning
     emmental_learner.learn(
-        mtl_model,
+        model,
         [train_dataloader, dev_dataloader],
     )
 
-    test_score = mtl_model.score(test_dataloader)
+    test_score = model.score(test_dataloader)
 
-    assert test_score["task1/synthetic/test/loss"] <= 0.1
+    assert test_score["task1/synthetic/test/loss"] <= 0.4
 
     shutil.rmtree(dirpath)