Fixing float infinity in attention-center GPU. (#1386)

Run inference-precision fp32. Overflow because of huge FC feature depths. Signed-off-by: hyunback <[email protected]>
openvinotoolkit · Oct 18, 2023 · 677c677 · 677c677
1 parent 702417e
commit 677c677
Showing 1 changed file with 2 additions and 0 deletions.
diff --git a/notebooks/216-attention-center/216-attention-center.ipynb b/notebooks/216-attention-center/216-attention-center.ipynb
@@ -184,6 +184,8 @@
    "metadata": {},
    "outputs": [],
    "source": [
+    "if \"GPU\" in device.value:\n",
+    "    core.set_property(device_name=device.value, properties={'INFERENCE_PRECISION_HINT': ov.Type.f32})\n",
     "compiled_model = core.compile_model(model=model, device_name=device.value)"
    ]
   },