Nomic Embed Text V2 Moe Msmarco Bpr
这是一个从nomic-ai/nomic-embed-text-v2-moe微调的sentence-transformers模型,可将文本映射到768维稠密向量空间,用于语义文本相似度计算等任务。
下载量 41
发布时间 : 3/4/2025
模型简介
该模型将句子和段落映射到768维稠密向量空间,可用于语义文本相似度计算、语义搜索、复述挖掘、文本分类、聚类等任务。
模型特点
长文本处理能力
支持最大8192个token的序列长度,适合处理长文本内容。
高效语义编码
将文本映射到768维稠密向量空间,保留丰富的语义信息。
微调优化
基于nomic-ai/nomic-embed-text-v2-moe模型进行微调,优化了语义相似度任务表现。
模型能力
语义文本相似度计算
语义搜索
复述挖掘
文本分类
文本聚类
使用案例
信息检索
相似问题匹配
在问答系统中匹配语义相似的问题
可准确识别不同表述但语义相同的问题
内容管理
文档去重
识别语义相似的文档内容
有效减少重复内容存储
🚀 基于nomic-ai/nomic-embed-text-v2-moe的句子转换器
这是一个基于 sentence-transformers 从 nomic-ai/nomic-embed-text-v2-moe 微调而来的模型。它能将句子和段落映射到768维的密集向量空间,可用于语义文本相似度计算、语义搜索、释义挖掘、文本分类、聚类等任务。
🚀 快速开始
本模型可以方便地应用于多种自然语言处理任务,以下是使用步骤和示例代码。
✨ 主要特性
- 高维映射:将句子和段落映射到768维的密集向量空间。
- 多任务适用:可用于语义文本相似度计算、语义搜索、释义挖掘、文本分类、聚类等。
- 长序列处理:最大序列长度可达8192个标记。
📦 安装指南
首先,你需要安装 Sentence Transformers 库:
pip install -U sentence-transformers
💻 使用示例
基础用法
from sentence_transformers import SentenceTransformer
# 从 🤗 Hub 下载模型
model = SentenceTransformer("BlackBeenie/nomic-embed-text-v2-moe-msmarco-bpr")
# 运行推理
sentences = [
'what services are offered by adult day care',
'Consumer Guide to Long Term Care. Adult Day Care. Adult day care is a planned program offered in a group setting that provides services that improve or maintain health or functioning, and social activities for seniors and persons with disabilities.',
'The Met Life Market survey of 2008 on adult day services states the average cost for adult day care services is $64 per day. There has been an increase of 5% in these services in the past year.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]
# 获取嵌入向量的相似度分数
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]
📚 详细文档
模型详情
模型描述
属性 | 详情 |
---|---|
模型类型 | 句子转换器 |
基础模型 | nomic-ai/nomic-embed-text-v2-moe |
最大序列长度 | 8192个标记 |
输出维度 | 768维 |
相似度函数 | 余弦相似度 |
模型来源
- 文档:Sentence Transformers 文档
- 仓库:GitHub 上的 Sentence Transformers
- Hugging Face:Hugging Face 上的 Sentence Transformers
完整模型架构
SentenceTransformer(
(0): Transformer({'max_seq_length': 8192, 'do_lower_case': False}) with Transformer model: NomicBertModel
(1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
)
训练详情
训练数据集
未命名数据集
- 规模:498,970 个训练样本
- 列名:
sentence_0
、sentence_1
和sentence_2
- 近似统计(基于前1000个样本):
sentence_0 sentence_1 sentence_2 类型 字符串 字符串 字符串 详情 - 最小:4个标记
- 平均:9.75个标记
- 最大:24个标记
- 最小:24个标记
- 平均:89.23个标记
- 最大:241个标记
- 最小:20个标记
- 平均:86.66个标记
- 最大:280个标记
- 样本:
sentence_0 sentence_1 sentence_2 what the history of bluetooth
When asked about the name Bluetooth, I explained that Bluetooth was borrowed from the 10th century, second King of Denmark, King Harald Bluetooth; who was famous for uniting Scandinavia just as we intended to unite the PC and cellular industries with a short-range wireless link.
Technology: 1 How secure is a Bluetooth network? 2 What is Frequency-Hopping Spread Spectrum (FHSS)? 3 Will other RF (Radio Frequency) devices interfere with Bluetooth Devices? 4 Will Bluetooth and Wireless LAN (WLAN) interfere with each other? 5 What is the data throughput speed of a Bluetooth connection? 6 What is the range of Bluetooth 7 ... What kind of ...
how thin can a concrete slab be
Another issue that must be addressed is the added weight of the thin-slab. Poured gypsum thin-slabs typically add 13 to 15 pounds per square foot to the dead loading of a floor structure. Standard weight concrete thin slabs add about 18 pounds per square foot (at 1.5 thickness).
Find the Area in square feet: We will use a concrete slab pour for our example. Letâs say that we need to figure out the yardage for a slab that will be 15 feet long by 10 feet wide and 4 inches thick. First we find the area by multiplying the length times the width. 1 15 feet X 10 feet = 150 square feet.
how long to cook eggs to hard boil
This method works best if the eggs are in a single layer, but you can double them up as well, you'll just need to add more time to the steaming time. 3 Set your timer for 6 minutes for soft boiled, 10 minutes for hard boiled with a still translucent and bright yolk, or 12-15 minutes for cooked-through hard boiled.
Hard-Steamed Eggs. Fill a pot that can comfortably hold your steamer with the lid on with 1 to 2 inches of water. Bring to a rolling boil, 212 degrees Fahrenheit. Place your eggs in a metal steamer, and lower the basket into the pot. The eggs should sit above the boiling water. Cover and cook for 12 minutes. Hard-steamed eggs, like hard-boiled eggs, are eggs that are cooked until the egg yolk is fully set and has turned to a chalky texture.
- 损失函数:
beir.losses.bpr_loss.BPRLoss
训练超参数
非默认超参数
eval_strategy
: stepsper_device_train_batch_size
: 32per_device_eval_batch_size
: 32num_train_epochs
: 5fp16
: Truemulti_dataset_batch_sampler
: round_robin
所有超参数
点击展开
overwrite_output_dir
: Falsedo_predict
: Falseeval_strategy
: stepsprediction_loss_only
: Trueper_device_train_batch_size
: 32per_device_eval_batch_size
: 32per_gpu_train_batch_size
: Noneper_gpu_eval_batch_size
: Nonegradient_accumulation_steps
: 1eval_accumulation_steps
: Nonetorch_empty_cache_steps
: Nonelearning_rate
: 5e-05weight_decay
: 0.0adam_beta1
: 0.9adam_beta2
: 0.999adam_epsilon
: 1e-08max_grad_norm
: 1num_train_epochs
: 5max_steps
: -1lr_scheduler_type
: linearlr_scheduler_kwargs
: {}warmup_ratio
: 0.0warmup_steps
: 0log_level
: passivelog_level_replica
: warninglog_on_each_node
: Truelogging_nan_inf_filter
: Truesave_safetensors
: Truesave_on_each_node
: Falsesave_only_model
: Falserestore_callback_states_from_checkpoint
: Falseno_cuda
: Falseuse_cpu
: Falseuse_mps_device
: Falseseed
: 42data_seed
: Nonejit_mode_eval
: Falseuse_ipex
: Falsebf16
: Falsefp16
: Truefp16_opt_level
: O1half_precision_backend
: autobf16_full_eval
: Falsefp16_full_eval
: Falsetf32
: Nonelocal_rank
: 0ddp_backend
: Nonetpu_num_cores
: Nonetpu_metrics_debug
: Falsedebug
: []dataloader_drop_last
: Falsedataloader_num_workers
: 0dataloader_prefetch_factor
: Nonepast_index
: -1disable_tqdm
: Falseremove_unused_columns
: Truelabel_names
: Noneload_best_model_at_end
: Falseignore_data_skip
: Falsefsdp
: []fsdp_min_num_params
: 0fsdp_config
: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}fsdp_transformer_layer_cls_to_wrap
: Noneaccelerator_config
: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}deepspeed
: Nonelabel_smoothing_factor
: 0.0optim
: adamw_torchoptim_args
: Noneadafactor
: Falsegroup_by_length
: Falselength_column_name
: lengthddp_find_unused_parameters
: Noneddp_bucket_cap_mb
: Noneddp_broadcast_buffers
: Falsedataloader_pin_memory
: Truedataloader_persistent_workers
: Falseskip_memory_metrics
: Trueuse_legacy_prediction_loop
: Falsepush_to_hub
: Falseresume_from_checkpoint
: Nonehub_model_id
: Nonehub_strategy
: every_savehub_private_repo
: Nonehub_always_push
: Falsegradient_checkpointing
: Falsegradient_checkpointing_kwargs
: Noneinclude_inputs_for_metrics
: Falseinclude_for_metrics
: []eval_do_concat_batches
: Truefp16_backend
: autopush_to_hub_model_id
: Nonepush_to_hub_organization
: Nonemp_parameters
:auto_find_batch_size
: Falsefull_determinism
: Falsetorchdynamo
: Noneray_scope
: lastddp_timeout
: 1800torch_compile
: Falsetorch_compile_backend
: Nonetorch_compile_mode
: Nonedispatch_batches
: Nonesplit_batches
: Noneinclude_tokens_per_second
: Falseinclude_num_input_tokens_seen
: Falseneftune_noise_alpha
: Noneoptim_target_modules
: Nonebatch_eval_metrics
: Falseeval_on_start
: Falseuse_liger_kernel
: Falseeval_use_gather_object
: Falseaverage_tokens_across_devices
: Falseprompts
: Nonebatch_sampler
: batch_samplermulti_dataset_batch_sampler
: round_robin
训练日志
点击展开
轮次 | 步数 | 训练损失 |
---|---|---|
0.0321 | 500 | 0.3396 |
0.0641 | 1000 | 0.2094 |
0.0962 | 1500 | 0.21 |
0.1283 | 2000 | 0.1955 |
0.1603 | 2500 | 0.1989 |
0.1924 | 3000 | 0.1851 |
0.2245 | 3500 | 0.1839 |
0.2565 | 4000 | 0.1859 |
0.2886 | 4500 | 0.1892 |
0.3207 | 5000 | 0.1865 |
0.3527 | 5500 | 0.1773 |
0.3848 | 6000 | 0.1796 |
0.4169 | 6500 | 0.1929 |
0.4489 | 7000 | 0.1829 |
0.4810 | 7500 | 0.172 |
0.5131 | 8000 | 0.1792 |
0.5451 | 8500 | 0.1747 |
0.5772 | 9000 | 0.1802 |
0.6092 | 9500 | 0.1856 |
0.6413 | 10000 | 0.1751 |
0.6734 | 10500 | 0.173 |
0.7054 | 11000 | 0.1774 |
0.7375 | 11500 | 0.1722 |
0.7696 | 12000 | 0.1825 |
0.8016 | 12500 | 0.1714 |
0.8337 | 13000 | 0.1732 |
0.8658 | 13500 | 0.167 |
0.8978 | 14000 | 0.1792 |
0.9299 | 14500 | 0.1697 |
0.9620 | 15000 | 0.1682 |
0.9940 | 15500 | 0.1764 |
1.0 | 15593 | - |
1.0261 | 16000 | 0.0875 |
1.0582 | 16500 | 0.0798 |
1.0902 | 17000 | 0.0764 |
1.1223 | 17500 | 0.0783 |
1.1544 | 18000 | 0.0759 |
1.1864 | 18500 | 0.0834 |
1.2185 | 19000 | 0.082 |
1.2506 | 19500 | 0.0827 |
1.2826 | 20000 | 0.0876 |
1.3147 | 20500 | 0.0819 |
1.3468 | 21000 | 0.0841 |
1.3788 | 21500 | 0.0815 |
1.4109 | 22000 | 0.0819 |
1.4430 | 22500 | 0.0883 |
1.4750 | 23000 | 0.0826 |
1.5071 | 23500 | 0.0837 |
1.5392 | 24000 | 0.086 |
1.5712 | 24500 | 0.0806 |
1.6033 | 25000 | 0.0918 |
1.6353 | 25500 | 0.0885 |
1.6674 | 26000 | 0.0885 |
1.6995 | 26500 | 0.088 |
1.7315 | 27000 | 0.0843 |
1.7636 | 27500 | 0.0915 |
1.7957 | 28000 | 0.0843 |
1.8277 | 28500 | 0.0868 |
1.8598 | 29000 | 0.0857 |
1.8919 | 29500 | 0.0931 |
1.9239 | 30000 | 0.0852 |
1.9560 | 30500 | 0.0913 |
1.9881 | 31000 | 0.0857 |
2.0 | 31186 | - |
2.0201 | 31500 | 0.0547 |
2.0522 | 32000 | 0.0459 |
2.0843 | 32500 | 0.0451 |
2.1163 | 33000 | 0.0407 |
2.1484 | 33500 | 0.0469 |
2.1805 | 34000 | 0.0459 |
2.2125 | 34500 | 0.0508 |
2.2446 | 35000 | 0.0508 |
2.2767 | 35500 | 0.0518 |
2.3087 | 36000 | 0.0552 |
2.3408 | 36500 | 0.0491 |
2.3729 | 37000 | 0.0575 |
2.4049 | 37500 | 0.0558 |
2.4370 | 38000 | 0.0475 |
2.4691 | 38500 | 0.0486 |
2.5011 | 39000 | 0.0536 |
2.5332 | 39500 | 0.0559 |
2.5653 | 40000 | 0.0524 |
2.5973 | 40500 | 0.0496 |
2.6294 | 41000 | 0.0486 |
2.6615 | 41500 | 0.0526 |
2.6935 | 42000 | 0.0443 |
2.7256 | 42500 | 0.058 |
2.7576 | 43000 | 0.0543 |
2.7897 | 43500 | 0.0527 |
2.8218 | 44000 | 0.0528 |
2.8538 | 44500 | 0.0573 |
2.8859 | 45000 | 0.0628 |
2.9180 | 45500 | 0.0443 |
2.9500 | 46000 | 0.0531 |
2.9821 | 46500 | 0.0554 |
3.0 | 46779 | - |
3.0142 | 47000 | 0.0346 |
3.0462 | 47500 | 0.0288 |
3.0783 | 48000 | 0.0219 |
3.1104 | 48500 | 0.0259 |
3.1424 | 49000 | 0.0237 |
3.1745 | 49500 | 0.0307 |
3.2066 | 50000 | 0.0234 |
3.2386 | 50500 | 0.0312 |
3.2707 | 51000 | 0.0297 |
3.3028 | 51500 | 0.0299 |
3.3348 | 52000 | 0.0326 |
3.3669 | 52500 | 0.0266 |
3.3990 | 53000 | 0.0296 |
3.4310 | 53500 | 0.0289 |
3.4631 | 54000 | 0.0216 |
3.4952 | 54500 | 0.0289 |
3.5272 | 55000 | 0.033 |
3.5593 | 55500 | 0.0248 |
3.5914 | 56000 | 0.0246 |
3.6234 | 56500 | 0.0287 |
3.6555 | 57000 | 0.0267 |
3.6876 | 57500 | 0.0285 |
3.7196 | 58000 | 0.0288 |
3.7517 | 58500 | 0.0283 |
3.7837 | 59000 | 0.0283 |
3.8158 | 59500 | 0.029 |
3.8479 | 60000 | 0.0327 |
3.8799 | 60500 | 0.0239 |
3.9120 | 61000 | 0.0356 |
3.9441 | 61500 | 0.0323 |
3.9761 | 62000 | 0.0213 |
4.0 | 62372 | - |
4.0082 | 62500 | 0.0275 |
4.0403 | 63000 | 0.0125 |
4.0723 | 63500 | 0.0183 |
4.1044 | 64000 | 0.0138 |
4.1365 | 64500 | 0.0174 |
4.1685 | 65000 | 0.0088 |
4.2006 | 65500 | 0.0126 |
4.2327 | 66000 | 0.0134 |
4.2647 | 66500 | 0.0099 |
4.2968 | 67000 | 0.0188 |
4.3289 | 67500 | 0.0112 |
4.3609 | 68000 | 0.0156 |
4.3930 | 68500 | 0.0175 |
4.4251 | 69000 | 0.0128 |
4.4571 | 69500 | 0.0154 |
4.4892 | 70000 | 0.0127 |
4.5213 | 70500 | 0.0131 |
4.5533 | 71000 | 0.017 |
4.5854 | 71500 | 0.0116 |
4.6175 | 72000 | 0.0137 |
4.6495 | 72500 | 0.0156 |
4.6816 | 73000 | 0.0155 |
4.7137 | 73500 | 0.0078 |
4.7457 | 74000 | 0.0152 |
4.7778 | 74500 | 0.0089 |
4.8099 | 75000 | 0.0116 |
4.8419 | 75500 | 0.0144 |
4.8740 | 76000 | 0.0112 |
4.9060 | 76500 | 0.0108 |
4.9381 | 77000 | 0.0188 |
4.9702 | 77500 | 0.0109 |
5.0 | 77965 | - |
框架版本
- Python: 3.11.11
- Sentence Transformers: 3.4.1
- Transformers: 4.49.0
- PyTorch: 2.5.1+cu124
- Accelerate: 1.3.0
- Datasets: 3.3.2
- Tokenizers: 0.21.0
📄 许可证
文档中未提及相关许可证信息。
📄 引用
BibTeX
Sentence Transformers
@inproceedings{reimers-2019-sentence-bert,
title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
author = "Reimers, Nils and Gurevych, Iryna",
booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
month = "11",
year = "2019",
publisher = "Association for Computational Linguistics",
url = "https://arxiv.org/abs/1908.10084",
}
Jina Embeddings V3
Jina Embeddings V3 是一个多语言句子嵌入模型,支持超过100种语言,专注于句子相似度和特征提取任务。
文本嵌入
Transformers 支持多种语言

J
jinaai
3.7M
911
Ms Marco MiniLM L6 V2
Apache-2.0
基于MS Marco段落排序任务训练的交叉编码器模型,用于信息检索中的查询-段落相关性评分
文本嵌入 英语
M
cross-encoder
2.5M
86
Opensearch Neural Sparse Encoding Doc V2 Distill
Apache-2.0
基于蒸馏技术的稀疏检索模型,专为OpenSearch优化,支持免推理文档编码,在搜索相关性和效率上优于V1版本
文本嵌入
Transformers 英语

O
opensearch-project
1.8M
7
Sapbert From PubMedBERT Fulltext
Apache-2.0
基于PubMedBERT的生物医学实体表征模型,通过自对齐预训练优化语义关系捕捉
文本嵌入 英语
S
cambridgeltl
1.7M
49
Gte Large
MIT
GTE-Large 是一个强大的句子转换器模型,专注于句子相似度和文本嵌入任务,在多个基准测试中表现出色。
文本嵌入 英语
G
thenlper
1.5M
278
Gte Base En V1.5
Apache-2.0
GTE-base-en-v1.5 是一个英文句子转换器模型,专注于句子相似度任务,在多个文本嵌入基准测试中表现优异。
文本嵌入
Transformers 支持多种语言

G
Alibaba-NLP
1.5M
63
Gte Multilingual Base
Apache-2.0
GTE Multilingual Base 是一个多语言的句子嵌入模型,支持超过50种语言,适用于句子相似度计算等任务。
文本嵌入
Transformers 支持多种语言

G
Alibaba-NLP
1.2M
246
Polybert
polyBERT是一个化学语言模型,旨在实现完全由机器驱动的超快聚合物信息学。它将PSMILES字符串映射为600维密集指纹,以数值形式表示聚合物化学结构。
文本嵌入
Transformers

P
kuelumbus
1.0M
5
Bert Base Turkish Cased Mean Nli Stsb Tr
Apache-2.0
基于土耳其语BERT的句子嵌入模型,专为语义相似度任务优化
文本嵌入
Transformers 其他

B
emrecan
1.0M
40
GIST Small Embedding V0
MIT
基于BAAI/bge-small-en-v1.5模型微调的文本嵌入模型,通过MEDI数据集与MTEB分类任务数据集训练,优化了检索任务的查询编码能力。
文本嵌入
Safetensors 英语
G
avsolatorio
945.68k
29
精选推荐AI模型
Llama 3 Typhoon V1.5x 8b Instruct
专为泰语设计的80亿参数指令模型,性能媲美GPT-3.5-turbo,优化了应用场景、检索增强生成、受限生成和推理任务
大型语言模型
Transformers 支持多种语言

L
scb10x
3,269
16
Cadet Tiny
Openrail
Cadet-Tiny是一个基于SODA数据集训练的超小型对话模型,专为边缘设备推理设计,体积仅为Cosmo-3B模型的2%左右。
对话系统
Transformers 英语

C
ToddGoldfarb
2,691
6
Roberta Base Chinese Extractive Qa
基于RoBERTa架构的中文抽取式问答模型,适用于从给定文本中提取答案的任务。
问答系统 中文
R
uer
2,694
98