---
title: sat-12l-sm-coreml
canonical_url: "https://www.modelscope.cn/models/ukk0708/sat-12l-sm-coreml"
md_url: "https://www.modelscope.cn/models/ukk0708/sat-12l-sm-coreml.md"
repository: ukk0708/sat-12l-sm-coreml
last_updated: 2026-10-24
license: mit
base_model:
  - segment-any-text/sat-12l-sm
base_model_relation: finetune
downloads: 0
stars: 0
tags:
  - sentence-segmentation
  - coreml
  - multilingual
---

# sat-12l-sm-coreml

> sat-12l-sm-coreml - ukk0708 在 ModelScope 开源的模型。sat-12l-sm for Core ML

ukk0708/sat-12l-sm-coreml 是 ModelScope 魔搭社区上的机器学习模型，采用 mit 许可，基于 segment-any-text/sat-12l-sm 构建。

- **Repository**: ukk0708/sat-12l-sm-coreml
- **License**: mit
- **Base model**: segment-any-text/sat-12l-sm
- **Tags**: sentence-segmentation, coreml, multilingual
- **Downloads**: 0
- **Stars**: 0
- **Last updated**: 2026-10-24

Source: https://www.modelscope.cn/models/ukk0708/sat-12l-sm-coreml

---

# sat-12l-sm for Core ML

A Core ML conversion of [segment-any-text/sat-12l-sm](https://huggingface.co/segment-any-text/sat-12l-sm)
(Segment any Text, [wtpsplit](https://github.com/segment-any-text/wtpsplit), MIT), used by the Vocena
app to find sentence ends in speech-recognition text that has no punctuation.

- `SaTSegmenter.mlmodelc`: compiled Core ML model. Inputs `input_ids` and `attention_mask`, Int32
  [1, N] with N ∈ {64, 128, 256, 512}; output `logits` Float32 [1, N], one sentence-end logit per token.
  Wrap each block as `<s>` (0) + tokens + `</s>` (2) and pad with `<pad>` (1), mask 0.
- Word and position embeddings are int8 (per-block, block 32); the 12 encoder layers stay fp16. On the
  validation texts (Thai, Chinese, English, Japanese) the cuts at thresholds 0.05 and 0.03 are identical
  to wtpsplit 2.2.2 with the original ONNX model. Runs on the CPU and Neural Engine
  (`MLComputeUnits.cpuAndNeuralEngine`).
- `tokenizer.json`: the [FacebookAI/xlm-roberta-base](https://huggingface.co/FacebookAI/xlm-roberta-base)
  tokenizer (MIT), unchanged. `tokenizer_config.json` names its class for swift-transformers.

Licence: MIT, as the original model and tokenizer.