Skip to content

armgit / laya

Public Imported from GitHub
0

Non-autoregressive System 1 decision engine. Typed choice, score and yes/no decisions over any text in a single forward pass, in 100+ languages, with a router that picks the right checkpoint per request.

docs(evals): keep exporter flags on lines that name the exporter

Adeshina Lasisi6df6351f2bd3
diff --git a/docs/evals.md b/docs/evals.md
index e7630332..d2051eaf 100644
--- a/docs/evals.md
+++ b/docs/evals.md
@@ -85,14 +85,15 @@ laya-evals run data.jsonl --onnx laya.int8.onnx --max-ece 0.05
```
The export **verifies itself before it exits**: it runs the graph it just wrote in ONNX Runtime at
-batch 1, 2 and 3 and compares against PyTorch, and with `--quantize` it does the same for the INT8
-copy at a looser tolerance. A graph that cannot serve more than one question fails the export instead
-of being written and deployed. That needs `onnxruntime` (`pip install laya[onnx]`); pass
-`--no-verify` to skip it if you only have the exporter installed. Marker width is dynamic in the exported
-graph for two or more markers (a graph traced at 2 runs at 3, 4, 5 and 8), but exactly one marker is
-baked out by a Python branch in the model, so a one-criterion `score` question is not servable from
-ONNX when every question in the batch has one criterion; use the torch path for those. There is no
-need to re-export per criteria count.
+batch 1, 2 and 3 and compares against PyTorch, and `scripts/export_onnx.py --quantize` does the same
+for the INT8 copy at a looser tolerance. A graph that cannot serve more than one question fails the
+export instead of being written and deployed. That needs `onnxruntime` (`pip install laya[onnx]`);
+`scripts/export_onnx.py --no-verify` skips it if you only have the exporter installed.
+
+Marker width is dynamic in the exported graph for two or more markers (a graph traced at 2 runs at
+3, 4, 5 and 8), but exactly one marker is baked out by a Python branch in the model, so a
+one-criterion `score` question is not servable from ONNX when every question in the batch has one
+criterion; use the torch path for those. There is no need to re-export per criteria count.
`--model` names the checkpoint the export came from — a Hub id or local path, not a Router
short name like `english`, since there is no Router on this path (default