opus-mt-sla-en Ascend NPU 部署指南

opus-mt-sla-en-ascend/
├── inference.py          # 推理测试脚本
├── log.txt               # 测试日志
├── README.md             # 本文档
├── test_sentences.txt    # 测试句子
└── precision_result.json # 精度测试结果

cd /data/ysws/agentsp/5-20-1/opus-mt-sla-en-ascend/

python3 inference.py

cd /data/ysws/agentsp/5-20-1/opus-mt-sla-en-ascend/

python3 inference.py --precision_test

============================================================
opus-mt-sla-en Ascend NPU 部署测试
============================================================
MODEL_DIR: /data/ysws/agentsp/5-20-1/Helsinki-NLP/opus-mt-sla-en
OUTPUT_DIR: /data/ysws/agentsp/5-20-1/opus-mt-sla-en-ascend
Mode: precision_test

============================================================
创建测试样本
============================================================
测试句子已保存到: /data/ysws/agentsp/5-20-1/opus-mt-sla-en-ascend/test_sentences.txt
共 8 句

============================================================
opus-mt-sla-en NPU 推理测试
============================================================
Device: npu:0
Model loaded successfully!

测试句子数量: 8
  [1] Dobry den, jak se mate?
  [2] Jake je pocasi?
  [3] Dekuji mockrat.
  [4] Na shledanou!
  [5] Kolik to stoji?
  [6] Nevim.
  [7] Jsem student.
  [8] Good morning!

开始翻译 (device: npu:0)...

翻译结果:
  [1] 原文: Dobry den, jak se mate?
      译文: Hello, how are you?
  [2] 原文: Jake je pocasi?
      译文: Jake's a pocha?
  [3] 原文: Dekuji mockrat.
      译文: Thank you very much.
  [4] 原文: Na shledanou!
      译文: Good-bye!
  [5] 原文: Kolik to stoji?
      译文: How much is it?
  [6] 原文: Nevim.
      译文: I don't know.
  [7] 原文: Jsem student.
      译文: I'm a student.
  [8] 原文: Good morning!
      译文: Good Morning!

总耗时: 1.7376s
平均每句: 0.2172s

============================================================
opus-mt-sla-en 精度测试 (CPU vs NPU)
============================================================
Device: npu:0

加载 CPU 模型...
CPU 模型加载完成

加载 NPU 模型...
NPU 模型加载完成

测试句子数量: 3

--- 句子 1 ---
原文: Dobry den, jak se mate?
CPU 译文: Hello, how are you?
CPU 耗时: 1.5959s
NPU 译文: Hello, how are you?
NPU 耗时: 0.1481s
译文匹配: True

--- 句子 2 ---
原文: Jake je pocasi?
CPU 译文: Jake's a pocha?
CPU 耗时: 1.8757s
NPU 译文: Jake's a pocha?
NPU 耗时: 0.1555s
译文匹配: True

--- 句子 3 ---
原文: Na shledanou!
CPU 译文: Good-bye!
CPU 耗时: 0.9935s
NPU 译文: Good-bye!
NPU 耗时: 0.0873s
译文匹配: True

============================================================
精度测试结果汇总
============================================================
译文完全匹配: PASS
平均 CPU 推理时间: 1.4883s
平均 NPU 推理时间: 0.1303s
NPU 加速比: 11.42x

精度阈值: 1.0%
译文匹配率: PASS

总体状态: PASS

============================================================
测试完成!
============================================================

import torch
from transformers import AutoModelForSeq2SeqLM, AutoTokenizer

MODEL_DIR = "/data/ysws/agentsp/5-20-1/Helsinki-NLP/opus-mt-sla-en"

tokenizer = AutoTokenizer.from_pretrained(MODEL_DIR)
model = AutoModelForSeq2SeqLM.from_pretrained(MODEL_DIR)
model = model.to("npu:0")
model.eval()

texts = ["Dobry den, jak se mate?"]
inputs = tokenizer(texts, return_tensors="pt", padding=True)
inputs = {k: v.to("npu:0") for k, v in inputs.items()}

with torch.no_grad():
    gen_ids = model.generate(
        inputs["input_ids"],
        attention_mask=inputs["attention_mask"],
        max_length=100,
        num_beams=4,
        early_stopping=True
    )

translations = tokenizer.batch_decode(gen_ids, skip_special_tokens=True)
print(translations)  # ['Hello, how are you?']