GIST-Embedding-v0 base trained on 380890 en-tr vacancy pairs
This is a sentence-transformers model finetuned from avsolatorio/GIST-Embedding-v0. It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.
Model Details
Model Description
- Model Type: Sentence Transformer
- Base model: avsolatorio/GIST-Embedding-v0
- Maximum Sequence Length: 512 tokens
- Output Dimensionality: 768 dimensions
- Similarity Function: Cosine Similarity
- Language: en
- License: mit
Model Sources
Full Model Architecture
SentenceTransformer(
(0): Transformer({'max_seq_length': 512, 'do_lower_case': True}) with Transformer model: BertModel
(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})
(2): Normalize()
)
Usage
Direct Usage (Sentence Transformers)
First install the Sentence Transformers library:
pip install -U sentence-transformers
Then you can load this model and run inference.
from sentence_transformers import SentenceTransformer
model = SentenceTransformer("avsolatorio/GIST-Embedding-v0_en-tr_all-jobs")
sentences = [
"RF-QK13895 Muhasebe Uzmanı MET-GÜN İNŞAAT TAAHHÜT VE TİCARET A.Ş. İstanbul(Avr.)\n\n**Şirketimizin\xa0 Maslak'ta Bulunan Genel Müdürlüğü'nde görevlendirilmek üzere ''Muhasebe Uzmanı'' arayışımız bulunmaktadır.**\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Üniversitelerin İlgili bölümlerinden mezun,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Alanında 4-5 yıl tecrübeli,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Tek düzen hesap planı ve vergi uygulamalarını bilen,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 E- defter, E-fatura, E-arşiv süreçlerine hakim,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Erp programları kullanmış,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Ofis programlarına iyi derecede hakim,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Analitik düşünme becerisi gelişmiş ve detaylara önem veren,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Takım çalışmasına yatkın, analiz ve değerlendirme yeteneğine sahip,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Erkek adaylar için askerlik görevini tamamlamış \n\n\n \n\n\nİŞ TANIMI\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Muhasebe kayıtlarına esas teşkil edecek belgelerin, şirket prosedürleri ve yasalar çerçevesinde kontrollerinin yapılması ve muhasebe sistemine kayıtlarının atılması,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 KDV, Muhtasar, Damga Vergisi işlemlerinin yapılması, \n\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Geçici ve Kurumlar Vergisi beyannamelerinin hazırlık sürecinde destek olunması,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 KDV İade süreçlerinin yürütülmesi,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Mizan kontrolü ve dönem sonu işlemlerinin yapılması,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Masraf formlarının takibi ve muhasebe kayıtlarının yapılması,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Cari hesap takibi ve mutabakatının yapılması,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Aylık Ba-Bs mutabakat formu çalışmalarının yapılması,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Muhasebesel raporlamalara destek olarak görev ve sorumluluk alınması,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Vergi işlemlerinin hazırlanması ve tahakkukların yapılması, ödeme işlemlerinin takibinin sağlanması\n\n \n\n\n",
'RF-QK13895 Accounting Specialist MET-GÜN CONSTRUCTION CONTRACTING AND TRADE INC. Istanbul(Avr.)\n\n**We are looking for an "Accounting Specialist" to be assigned to our company\'s General Directorate in Maslak.**\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Graduated from the relevant departments of universities,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 4-5 years of experience in the field,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Knowing the uniform chart of accounts and tax practices,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Proficient in e-ledger, e-invoice, e-archive processes,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Used ERP programs,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Good command of office programs,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Analytical thinking skills and attaching importance to details,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Prone to teamwork, ability to analyze and evaluate,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Completed military service for male candidates \n\nJOB DESCRIPTION\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Checking the documents that will form the basis of the accounting records within the framework of company procedures and laws and recording them in the accounting system,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Carrying out VAT, Withholding and Stamp Duty transactions, \n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Supporting the preparation process of Provisional and Corporate Tax returns,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Execution of VAT Refund processes,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Trial balance control and end-of-period transactions,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Follow-up of expense forms and accounting records,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Follow-up and reconciliation of current accounts,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Carrying out monthly Ba-Bs reconciliation form studies,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Taking duties and responsibilities by supporting accounting reporting,\n\n·\xa0\xa0\xa0\xa0\xa0\xa0\xa0 Preparation of tax transactions and accruals, follow-up of payment transactions\n\n',
'RF-HD45885 German Teacher Teleperformance Uşak <h2>GENERAL QUALIFICATIONS AND JOB DESCRIPTION</h2><h3><p class="x_x_x_x_x_x_x_x_x_x_x_x_xxxxxxxxxxxmsonormal"><span>Teleperformance, which is a leader in providing excellent customer experience worldwide, provides services to companies on an international scale, achieving superior results in the management of customer service, technical support, sales, marketing and collection processes of institutions. </span><span></span></p><p><span>Teleperformance Group, With more than <span>420,000</span> employees in <span>88</span> countries, <span>265+</span> languages and dialects, <span>and 450</span> Call centers , it provides outsourced Call Center services to large-scale international companies operating in different sectors and various countries.</span></p><p><span> </span></p><p class="x_x_x_x_x_x_x_x_x_x_x_x_xxxxxxxxxxxmsonormal">We are <span><span>looking for</span> a <span>"German Teacher" </span><span>to work at our Teleperformance Uşak Campus</span>.</span></p><p class="x_x_x_x_x_x_x_x_x_x_x_x_xxxxxxxxxxxmsonormal"><span> </span></p><ul><li><span>Graduated from German Language Teaching, Translation and Interpreting, German Language and Literature departments of universities,</span></li> <li><span>Having vocational education certificates,</span></li> At <li>least 3 years of school and language course instructor experience,</li> <li><span>Able to teach full-time to adults at General German (A1-B2) levels,</span></li><li><span>Strong human relations, high energy,</span></li> <li><span>prone to teamwork,</span></li> <li>no travel barriers between locations.</li></ul><p class="x_x_x_x_x_x_x_x_x_x_x_x_xxxxxxxxxxxmsonormal"><span> </span></p><p class="x_x_x_x_x_x_x_x_x_x_x_x_xxxxxxxxxxxmsonormal"><span>Job Description</span></p> Determining <ul><li><span>the level for the German classes to be opened in the relevant campuses of Teleperformance Turkey, creating the relevant curricula,</span></li> following <li><span>up the attendance of the participants in the German classes,</span></li><li><span>Organization of midterm exams and exam processes of the participants and reporting the results,</span></li> <li><span>Keeping the education provided alive with up-to-date resource researches,</span></li> <li><span>Preparing reports of all reports and processes related to the training classes and transferring them to the senior management.</span></li></ul></h3> ',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
Evaluation
Metrics
Triplet
- Datasets:
en-tr-jobs-validation and en-tr-jobs-test
- Evaluated with
TripletEvaluator
| Metric |
en-tr-jobs-validation |
en-tr-jobs-test |
| cosine_accuracy |
1.0 |
1.0 |
Training Details
Training Dataset
Unnamed Dataset
Evaluation Dataset
Unnamed Dataset
Training Hyperparameters
Non-Default Hyperparameters
eval_strategy: steps
per_device_train_batch_size: 64
per_device_eval_batch_size: 128
learning_rate: 2e-05
num_train_epochs: 5
warmup_ratio: 0.1
bf16: True
load_best_model_at_end: True
push_to_hub: True
hub_private_repo: True
batch_sampler: no_duplicates
All Hyperparameters
Click to expand
overwrite_output_dir: False
do_predict: False
eval_strategy: steps
prediction_loss_only: True
per_device_train_batch_size: 64
per_device_eval_batch_size: 128
per_gpu_train_batch_size: None
per_gpu_eval_batch_size: None
gradient_accumulation_steps: 1
eval_accumulation_steps: None
torch_empty_cache_steps: None
learning_rate: 2e-05
weight_decay: 0.0
adam_beta1: 0.9
adam_beta2: 0.999
adam_epsilon: 1e-08
max_grad_norm: 1.0
num_train_epochs: 5
max_steps: -1
lr_scheduler_type: linear
lr_scheduler_kwargs: {}
warmup_ratio: 0.1
warmup_steps: 0
log_level: passive
log_level_replica: warning
log_on_each_node: True
logging_nan_inf_filter: True
save_safetensors: True
save_on_each_node: False
save_only_model: False
restore_callback_states_from_checkpoint: False
no_cuda: False
use_cpu: False
use_mps_device: False
seed: 42
data_seed: None
jit_mode_eval: False
use_ipex: False
bf16: True
fp16: False
fp16_opt_level: O1
half_precision_backend: auto
bf16_full_eval: False
fp16_full_eval: False
tf32: None
local_rank: 0
ddp_backend: None
tpu_num_cores: None
tpu_metrics_debug: False
debug: []
dataloader_drop_last: False
dataloader_num_workers: 0
dataloader_prefetch_factor: None
past_index: -1
disable_tqdm: False
remove_unused_columns: True
label_names: None
load_best_model_at_end: True
ignore_data_skip: False
fsdp: []
fsdp_min_num_params: 0
fsdp_config: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
fsdp_transformer_layer_cls_to_wrap: None
accelerator_config: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
deepspeed: None
label_smoothing_factor: 0.0
optim: adamw_torch
optim_args: None
adafactor: False
group_by_length: False
length_column_name: length
ddp_find_unused_parameters: None
ddp_bucket_cap_mb: None
ddp_broadcast_buffers: False
dataloader_pin_memory: True
dataloader_persistent_workers: False
skip_memory_metrics: True
use_legacy_prediction_loop: False
push_to_hub: True
resume_from_checkpoint: None
hub_model_id: None
hub_strategy: every_save
hub_private_repo: True
hub_always_push: False
gradient_checkpointing: False
gradient_checkpointing_kwargs: None
include_inputs_for_metrics: False
include_for_metrics: []
eval_do_concat_batches: True
fp16_backend: auto
push_to_hub_model_id: None
push_to_hub_organization: None
mp_parameters:
auto_find_batch_size: False
full_determinism: False
torchdynamo: None
ray_scope: last
ddp_timeout: 1800
torch_compile: False
torch_compile_backend: None
torch_compile_mode: None
dispatch_batches: None
split_batches: None
include_tokens_per_second: False
include_num_input_tokens_seen: False
neftune_noise_alpha: None
optim_target_modules: None
batch_eval_metrics: False
eval_on_start: False
use_liger_kernel: False
eval_use_gather_object: False
average_tokens_across_devices: False
prompts: None
batch_sampler: no_duplicates
multi_dataset_batch_sampler: proportional
Training Logs
| Epoch |
Step |
Training Loss |
Validation Loss |
en-tr-jobs-validation_cosine_accuracy |
en-tr-jobs-test_cosine_accuracy |
| -1 |
-1 |
- |
- |
0.9833 |
- |
| 0.1680 |
500 |
0.7731 |
0.0033 |
1.0 |
- |
| 0.3360 |
1000 |
0.1047 |
0.0015 |
1.0 |
- |
| 0.5040 |
1500 |
0.0545 |
0.0009 |
1.0 |
- |
| 0.6720 |
2000 |
0.0388 |
0.0002 |
1.0 |
- |
| 0.8401 |
2500 |
0.03 |
0.0004 |
1.0 |
- |
| 1.0077 |
3000 |
0.0239 |
0.0004 |
1.0 |
- |
| 1.1757 |
3500 |
0.0201 |
0.0004 |
1.0 |
- |
| 1.3438 |
4000 |
0.0188 |
0.0002 |
1.0 |
- |
| 1.5118 |
4500 |
0.0189 |
0.0003 |
1.0 |
- |
| 1.6798 |
5000 |
0.0183 |
0.0001 |
1.0 |
- |
| 1.8478 |
5500 |
0.0163 |
0.0002 |
1.0 |
- |
| 2.0158 |
6000 |
0.0151 |
0.0001 |
1.0 |
- |
| 2.1838 |
6500 |
0.0135 |
0.0003 |
1.0 |
- |
| 2.3518 |
7000 |
0.0122 |
0.0002 |
1.0 |
- |
| 2.5198 |
7500 |
0.0127 |
0.0002 |
1.0 |
- |
| 2.6878 |
8000 |
0.0113 |
0.0001 |
1.0 |
- |
| 2.8558 |
8500 |
0.0113 |
0.0001 |
1.0 |
- |
| 3.0239 |
9000 |
0.0104 |
0.0002 |
1.0 |
- |
| 3.1919 |
9500 |
0.0104 |
0.0002 |
1.0 |
- |
| 3.3599 |
10000 |
0.0088 |
0.0002 |
1.0 |
- |
| 3.5279 |
10500 |
0.0108 |
0.0003 |
1.0 |
- |
| 3.6959 |
11000 |
0.0108 |
0.0001 |
1.0 |
- |
| 3.8639 |
11500 |
0.0089 |
0.0001 |
1.0 |
- |
| 4.0316 |
12000 |
0.0085 |
0.0001 |
1.0 |
- |
| 4.1996 |
12500 |
0.0091 |
0.0001 |
1.0 |
- |
| 4.3676 |
13000 |
0.0089 |
0.0001 |
1.0 |
- |
| 4.5356 |
13500 |
0.0098 |
0.0001 |
1.0 |
- |
| 4.7036 |
14000 |
0.0095 |
0.0001 |
1.0 |
- |
| 4.8716 |
14500 |
0.009 |
0.0001 |
1.0 |
- |
| -1 |
-1 |
- |
- |
- |
1.0 |
- The bold row denotes the saved checkpoint.
Framework Versions
- Python: 3.10.10
- Sentence Transformers: 3.4.1
- Transformers: 4.48.3
- PyTorch: 2.6.0+cu124
- Accelerate: 1.3.0
- Datasets: 3.3.0
- Tokenizers: 0.21.0
Citation
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",
}
MultipleNegativesRankingLoss
@misc{henderson2017efficient,
title={Efficient Natural Language Response Suggestion for Smart Reply},
author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil},
year={2017},
eprint={1705.00652},
archivePrefix={arXiv},
primaryClass={cs.CL}
}