Downloads · 30 days
0
jiangzhuo9357/sat-3l-sm-onnx
sat-3l-sm-onnx is a machine learning model from jiangzhuo9357. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as mit.
Weight-only 8-bit ONNX export of SaT 3L-SM (segment-any-text/sat-3l-sm), wtpsplit's 3-layer XLM-RoBERTa sentence-boundary model, with an 8-bit GatherBlockQuantized word-embedding table (the "q8w-gather" build), for us…
Downloads · 30 days
0
Access
Public
Updated Sep 18, 2026
Repo size
242 MB
Likes
0
Public
Click a slice to open those files.
.onnx242 MB · 96%
From the Hugging Face model README
Weight-only 8-bit ONNX export of SaT 3L-SM (segment-any-text/sat-3l-sm), wtpsplit's
3-layer XLM-RoBERTa sentence-boundary model, with an 8-bit GatherBlockQuantized word-embedding
table (the "q8w-gather" build), for use by
Sokuji's local sentence-segmentation stage.
LICENSE in this repo is the verbatim upstream text).| file | bytes | sha256 |
|---|---|---|
model.onnx | 241,945,842 | 1132546a16fa1aa9da218602049c6e7aeba0ec5b76409131446f40ebda2be99f |
tokenizer.json | 9,096,718 | a898ea75433890f6610f4e470b8ebeb0c21dce5c8dd61f892eb09eb5919d2e2c |
tokenizer.json is FacebookAI/xlm-roberta-base's tokenizer, unmodified.
The published model is fp16 and fails to load on WebGPU adapters without shader-f16 support
("Program Gather requires f16 but the device does not support it"). This build instead:
FLOAT16 initializer, Constant,
ConstantOfShape value, I/O type and Cast target converted to FLOAT, and a missing
com.microsoft opset import added so onnx.checker accepts the graph — the graph uses
SkipLayerNormalization and BiasGelu from that domain without declaring it).com.microsoft:MatMulNBits, 8 bits, block
size 32, symmetric (onnxruntime.quantization.matmul_nbits_quantizer.MatMulNBitsQuantizer) —
the same recipe as fireredpunc-onnx's weight-only build.Gather with com.microsoft:GatherBlockQuantized,
8 bits, block size 32 along the hidden axis: uint8 data with the implicit zero point 128 and
one float32 scale per 32-value block (max|w| / 127), built by hand because
onnxruntime's own quantizer only emits a 4-bit GatherBlockQuantized.No tensor in the result is float16, so it needs no shader-f16 support and runs on both the
WebGPU and WASM execution providers of onnxruntime-web 1.26.
Exact command, run against Sokuji's punctuation benchmark harness
(benchmark/punctuation-restoration/):
python parity/sat-3l-sm-q8w.py --force
This writes both sat/sat-3l-sm-q8w/model.onnx (the MatMulNBits-only build, 793,953,242 bytes,
word embedding left float32) and sat/sat-3l-sm-q8w-gather/model.onnx (the same plus the 8-bit
GatherBlockQuantized embedding, 241,945,842 bytes) plus a copied tokenizer.json in each
directory. model.onnx here is the q8w-gather output, copied unmodified — the smaller and
lower-memory of the two, and the one recommended for shipping.
From docs/superpowers/notes/2026-09-14-asr-punctuation-benchmark.md (Sokuji repo, 2026-09-14):
q8w-gather reproduces 261/270 (96.7%) identical
splits; every flip has both boundary probabilities within 0.015 of the 0.25 decision
threshold. The fp16 file itself, run on WASM, upcasts to float32 at load and matches a true
fp32 rewrite on all 270 rows.MIT, carried over from segment-any-text/sat-3l-sm and segment-any-text/wtpsplit. See
LICENSE in this repository (verbatim copy of the upstream wtpsplit LICENSE file).