compiled artefacts from mac_arm64 (cpu)
Browse files
README.md
CHANGED
|
@@ -258,9 +258,10 @@ Flat layout, named per **runtime × GPU arch × OS** (single-profile — no leng
|
|
| 258 |
`XLMRobertaForTokenClassification` layout with the 17 labels in the config:
|
| 259 |
`AutoModelForTokenClassification` loads them with no missing or unexpected keys, and its
|
| 260 |
per-token logits match the ONNX to 1e-5. Calibrate them with `temperature.json` exactly as
|
| 261 |
-
in the snippet above.
|
| 262 |
|
| 263 |
-
|
|
|
|
| 264 |
|
| 265 |
## License
|
| 266 |
|
|
|
|
| 258 |
`XLMRobertaForTokenClassification` layout with the 17 labels in the config:
|
| 259 |
`AutoModelForTokenClassification` loads them with no missing or unexpected keys, and its
|
| 260 |
per-token logits match the ONNX to 1e-5. Calibrate them with `temperature.json` exactly as
|
| 261 |
+
in the snippet above. The Apple (MLX) build is prepared from this pair.
|
| 262 |
|
| 263 |
+
On Apple Silicon the daemon runs this model through MLX (`model_gpu_mlx{VER}/`, prepared from
|
| 264 |
+
the raw weights above), the only GPU lane there.
|
| 265 |
|
| 266 |
## License
|
| 267 |
|