diff --git a/docs/source/reference/configs/training/training-config.md b/docs/source/reference/configs/training/training-config.md
index 5f843cf..bbc1ebb 100644
--- a/docs/source/reference/configs/training/training-config.md
+++ b/docs/source/reference/configs/training/training-config.md
@@ -9,8 +9,6 @@ Defined in `batdetect2.train.config`.
 
 - `compile_model`
   - compile the detector before training starts. This is off by default.
-- `precision`
-  - optional float32 matrix multiplication precision setting passed to PyTorch.
 - `train_loader`
   - training data loading and clipping settings.
 - `val_loader`
@@ -37,8 +35,7 @@ Use `TrainingConfig` when you want to change things like:
 - batch size,
 - augmentation,
 - optimiser and scheduler settings,
-- runtime options such as model compilation and matrix multiplication
-  precision,
+- runtime options such as model compilation,
 - number of epochs,
 - validation frequency,
 - checkpoint behaviour.
@@ -50,9 +47,6 @@ training. This can help on longer runs with stable tensor shapes, but it may be
 slower for short CPU-only experiments because PyTorch has to compile the graph
 before it can reuse it.
 
-Use `precision` to set PyTorch's float32 matrix multiplication precision before
-training starts. Supported values are `medium` and `high`.
-
 Example files live under `example_data/configs/`, including
 `example_data/configs/training.yaml`.
 
diff --git a/src/batdetect2/train/config.py b/src/batdetect2/train/config.py
index 8aec47d..ed1e201 100644
--- a/src/batdetect2/train/config.py
+++ b/src/batdetect2/train/config.py
@@ -1,5 +1,3 @@
-from typing import Literal
-
 from pydantic import Field
 
 from batdetect2.core.configs import BaseConfig
@@ -42,7 +40,6 @@ class PLTrainerConfig(BaseConfig):
 
 class TrainingConfig(BaseConfig):
     compile_model: bool = False
-    precision: Literal["medium", "high"] | None = None
     train_loader: TrainLoaderConfig = Field(default_factory=TrainLoaderConfig)
     val_loader: ValLoaderConfig = Field(default_factory=ValLoaderConfig)
     optimizer: OptimizerConfig = Field(default_factory=AdamOptimizerConfig)
diff --git a/src/batdetect2/train/train.py b/src/batdetect2/train/train.py
index f9430c3..e36c7ac 100644
--- a/src/batdetect2/train/train.py
+++ b/src/batdetect2/train/train.py
@@ -2,7 +2,6 @@ from collections.abc import Sequence
 from pathlib import Path
 from typing import Optional
 
-import torch
 from lightning import Trainer, seed_everything
 from lightning.pytorch.loggers import Logger
 from loguru import logger
@@ -208,13 +207,6 @@ def run_train(
         run_name=run_name,
     )
 
-    if train_config.precision is not None:
-        logger.info(
-            "Setting float32 matmul precision to {}",
-            train_config.precision,
-        )
-        torch.set_float32_matmul_precision(train_config.precision)
-
     if train_config.compile_model:
         logger.info("Compiling detector...")
         compile_model(module.model)
diff --git a/tests/test_train/test_lightning.py b/tests/test_train/test_lightning.py
index feb9a97..35e8edc 100644
--- a/tests/test_train/test_lightning.py
+++ b/tests/test_train/test_lightning.py
@@ -349,33 +349,6 @@ def test_run_train_compiles_detector_when_train_config_requests_compile(
     assert recorder.call_count > 0
 
 
-@pytest.mark.slow
-def test_run_train_sets_float32_matmul_precision(
-    tmp_path: Path,
-    example_annotations: list[data.ClipAnnotation],
-) -> None:
-    original_precision = torch.get_float32_matmul_precision()
-    train_config = build_fast_train_config()
-    train_config.precision = "high"
-
-    try:
-        run_train(
-            train_annotations=example_annotations[:1],
-            val_annotations=example_annotations[:1],
-            train_config=train_config,
-            num_epochs=1,
-            train_workers=0,
-            val_workers=0,
-            checkpoint_dir=tmp_path / "checkpoints",
-            log_dir=tmp_path / "logs",
-            seed=0,
-        )
-
-        assert torch.get_float32_matmul_precision() == "high"
-    finally:
-        torch.set_float32_matmul_precision(original_precision)
-
-
 def test_build_training_module_uses_provided_model() -> None:
     targets = build_targets(TargetConfig())
     roi_mapper = build_roi_mapping(TargetConfig().roi)