Instructions to use yennguyen45/moxhimt-60-web with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers.js
How to use yennguyen45/moxhimt-60-web with Transformers.js:
// npm i @huggingface/transformers import { pipeline } from '@huggingface/transformers'; // Allocate pipeline const pipe = await pipeline('translation', 'yennguyen45/moxhimt-60-web');
moxhimt-60-web (ONNX, INT8, web build)
Browser-ready build of DanVP/MoxhiMT-60, a MarianMT model for Chinese -> Vietnamese web-novel translation. Webnovel/xianxia, 57M.
All credit for training the base model goes to the original author (DanVP). This repo only adds:
- ONNX export with KV-cache (
encoder_model_quantized.onnx+decoder_model_merged_quantized.onnx) - Dynamic INT8 quantization (~4x smaller than fp32)
- A hand-built
tokenizer.jsonfast-tokenizer (the original model only ships slow sentencepiece files, whichtransformers.jscannot load directly) — verified to produce byte-identical token ids to the original slowMarianTokenizer.
Usage (transformers.js)
import { pipeline } from '@huggingface/transformers';
const translator = await pipeline('translation', 'yennguyen45/moxhimt-60-web', { dtype: 'q8' });
const out = await translator('你确定要这么做吗?');
console.log(out[0].translation_text);
License
CC-BY-4.0, inherited from the base model. Please credit the original author (DanVP) when using this model.
- Downloads last month
- 5
Model tree for yennguyen45/moxhimt-60-web
Base model
DanVP/MoxhiMT-60