xlm-roberta-longformer-base-16384

⚠️ 这只是 hyperonym/xlm-roberta-longformer-base-16384 的 PyTorch 版本，未做任何修改。

xlm-roberta-longformer 是一个多语言 Longformer，使用 XLM-RoBERTa 的权重进行初始化，未经过进一步预训练。该模型旨在针对下游任务进行微调。

用于复现模型的笔记本可在 GitHub 上获取：https://github.com/hyperonym/dirge/blob/master/models/xlm-roberta-longformer/convert.ipynb

在 Openmind 中使用

from openmind import AutoTokenizer, AutoModelForSequenceClassification, is_torch_npu_available
from openmind_hub import snapshot_download
import torch.nn.functional as F
from torch import Tensor
import openmind
import torch
import argparse
import time

# Mean Pooling - Take attention mask into account for correct averaging
def mean_pooling(model_output, attention_mask):
    token_embeddings = model_output[0]  # First element of model_output contains all token embeddings
    input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float()
    return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9)

def parse_args():
    parser = argparse.ArgumentParser()
    parser.add_argument(
        "--model_name_or_path",
        type=str,
        help="Path to model",
        default="jeffding/xlm-roberta-longformer-base-16384-openmind",
    )
    args = parser.parse_args()
    return args

def main():
    args = parse_args()
    model_path = args.model_name_or_path

    if is_torch_npu_available():
        device = "npu:0"
    else:
        device = "cpu"
        
    # Load model from HuggingFace Hub
    tokenizer = AutoTokenizer.from_pretrained(model_path)
    model = AutoModelForSequenceClassification.from_pretrained(
        model_path, trust_remote_code=True,
        torch_dtype=torch.float16
    ).to(device)
    model.eval()

    start_time = time.time()
    
    pairs = [["中国的首都在哪儿","北京"], ["what is the capital of China?", "北京"],["how to implement quick sort in python?","Introduction of quick sort"]]
    
    with torch.no_grad():
        inputs = tokenizer(pairs, padding=True, truncation=True, return_tensors='pt', max_length=512).to(device)
        scores = model(**inputs, return_dict=True).logits.view(-1, ).float()
        print(scores)
    
    end_time = time.time()
    print(f"硬件环境：{device},推理执行时间：{end_time - start_time}秒")
    
if __name__ == "__main__":
    main()

xlm-roberta-longformer-base-16384

xlm-roberta-longformer 是一个多语言 Longformer，使用 XLM-RoBERTa 的权重进行初始化，未经过进一步预训练。该模型旨在针对下游任务进行微调。

在 Openmind 中使用

from openmind import AutoTokenizer, AutoModelForSequenceClassification, is_torch_npu_available
from openmind_hub import snapshot_download
import torch.nn.functional as F
from torch import Tensor
import openmind
import torch
import argparse
import time

# Mean Pooling - Take attention mask into account for correct averaging
def mean_pooling(model_output, attention_mask):
    token_embeddings = model_output[0]  # First element of model_output contains all token embeddings
    input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float()
    return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9)

def parse_args():
    parser = argparse.ArgumentParser()
    parser.add_argument(
        "--model_name_or_path",
        type=str,
        help="Path to model",
        default="jeffding/xlm-roberta-longformer-base-16384-openmind",
    )
    args = parser.parse_args()
    return args

def main():
    args = parse_args()
    model_path = args.model_name_or_path

    if is_torch_npu_available():
        device = "npu:0"
    else:
        device = "cpu"
        
    # Load model from HuggingFace Hub
    tokenizer = AutoTokenizer.from_pretrained(model_path)
    model = AutoModelForSequenceClassification.from_pretrained(
        model_path, trust_remote_code=True,
        torch_dtype=torch.float16
    ).to(device)
    model.eval()

    start_time = time.time()
    
    pairs = [["中国的首都在哪儿","北京"], ["what is the capital of China?", "北京"],["how to implement quick sort in python?","Introduction of quick sort"]]
    
    with torch.no_grad():
        inputs = tokenizer(pairs, padding=True, truncation=True, return_tensors='pt', max_length=512).to(device)
        scores = model(**inputs, return_dict=True).logits.view(-1, ).float()
        print(scores)
    
    end_time = time.time()
    print(f"硬件环境：{device},推理执行时间：{end_time - start_time}秒")
    
if __name__ == "__main__":
    main()