Quantizer supported by the Arm Vgf backend. .. warning:: Setting ``use_composable_quantizer=True`` enables an experimental API surface that may change without notice. Args: compile_spec (VgfCompileSpec): Backend compile specification for Vgf targets.
| 1258 | |
| 1259 | |
| 1260 | class VgfQuantizer(TOSAQuantizer): |
| 1261 | """Quantizer supported by the Arm Vgf backend. |
| 1262 | |
| 1263 | .. warning:: |
| 1264 | Setting ``use_composable_quantizer=True`` enables an experimental API |
| 1265 | surface that may change without notice. |
| 1266 | |
| 1267 | Args: |
| 1268 | compile_spec (VgfCompileSpec): Backend compile specification for Vgf |
| 1269 | targets. |
| 1270 | use_composable_quantizer (bool): Whether to use the composable quantizer implementation. See https://github.com/pytorch/executorch/issues/17701" for details. |
| 1271 | |
| 1272 | """ |
| 1273 | |
| 1274 | def __init__( |
| 1275 | self, |
| 1276 | compile_spec: VgfCompileSpec, |
| 1277 | use_composable_quantizer: bool = False, |
| 1278 | ) -> None: |
| 1279 | super().__init__(compile_spec, use_composable_quantizer) |
no outgoing calls