test: update lat benchmark test

Author: okunator

cellseg_models_pytorch/utils/tests/test_latency_benchmarker.py

@@ -1,32 +1,32 @@
-# from cellseg_models_pytorch.inference import ResizeInferer
-# from cellseg_models_pytorch.models import cellpose_plus
-# from cellseg_models_pytorch.utils.latency_benchmark import LatencyBenchmarker
+from cellseg_models_pytorch.inference import ResizeInferer
+from cellseg_models_pytorch.models import cellpose_plus
+from cellseg_models_pytorch.utils.latency_benchmark import LatencyBenchmarker
 
 
-# def test_latency_benchmark(img_dir):
-#     model = cellpose_plus(sem_classes=3, type_classes=3, long_skip="unet")
+def test_latency_benchmark(img_dir):
+    model = cellpose_plus(sem_classes=3, type_classes=3, long_skip="unet")
 
-#     inferer = ResizeInferer(
-#         model,
-#         img_dir,
-#         out_activations={"sem": "softmax", "type": "softmax", "cellpose": "tanh"},
-#         out_boundary_weights={"sem": False, "type": False, "cellpose": True},
-#         resize=(256, 256),
-#         padding=80,
-#         instance_postproc="hovernet",
-#         batch_size=1,
-#         save_intermediate=True,
-#         device="cpu",
-#         parallel=False,
-#     )
-#     inferer.infer()
+    inferer = ResizeInferer(
+        model,
+        img_dir,
+        out_activations={"sem": "softmax", "type": "softmax", "cellpose": "tanh"},
+        out_boundary_weights={"sem": False, "type": False, "cellpose": True},
+        resize=(256, 256),
+        padding=80,
+        instance_postproc="hovernet",
+        batch_size=1,
+        save_intermediate=True,
+        device="cpu",
+        parallel=False,
+    )
+    inferer.infer()
 
-#     bm = LatencyBenchmarker(inferer)
+    bm = LatencyBenchmarker(inferer)
 
-#     bm.postproc_latency("inst", reps_per_img=1)
-#     bm.postproc_latency("type", reps_per_img=1)
-#     bm.postproc_latency("sem", reps_per_img=1)
-#     bm.inference_latency(reps=1, warmup_reps=0)
-#     bm.inference_postproc_latency(reps=1)
-#     # bm.model_latency(input_size=(64, 64), reps=1, warmup_reps=0, device="cpu")
-#     # bm.model_throughput(input_size=(64, 64), reps=1, warmup_reps=0, device="cpu")
+    bm.postproc_latency("inst", reps_per_img=1)
+    bm.postproc_latency("type", reps_per_img=1)
+    bm.postproc_latency("sem", reps_per_img=1)
+    bm.inference_latency(reps=1, warmup_reps=0)
+    bm.inference_postproc_latency(reps=1)
+    # bm.model_latency(input_size=(64, 64), reps=1, warmup_reps=0, device="cpu")
+    # bm.model_throughput(input_size=(64, 64), reps=1, warmup_reps=0, device="cpu")