From cf7d8169ec12cb3adf90d4ff78b3257e74985302 Mon Sep 17 00:00:00 2001 From: Onur Gündüz Date: Tue, 29 Sep 2026 13:05:44 +0200 Subject: Remove unused model families and inference paths --- README.md | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'README.md') diff --git a/README.md b/README.md index 2fe792b..a7f5aae 100644 --- a/README.md +++ b/README.md @@ -38,7 +38,9 @@ Use `--` before text that begins with a hyphen. Diagnostics go to stderr. ## Inference `langdetect.c` loads and validates the four F32 classifier tensors and the -`id2label` mapping. The encoder uses first-token pooling without L2 normalization. +`id2label` mapping. The encoder accepts F32 XLM-R weights with exact GeLU activation; +other model families and reduced-precision weights are not supported. +The encoder uses first-token pooling without L2 normalization. The classifier applies a 768-to-768 linear layer, `tanh`, a 768-to-20 linear layer, and stable softmax. Dropout is disabled during inference. -- cgit v1.3.1