SentenceTransformer based on TechWolf/JobBERT-v3

This is a sentence-transformers model finetuned from TechWolf/JobBERT-v3. It maps sentences & paragraphs to a 1024-dimensional dense vector space and can be used for retrieval.

Model Details

Model Description

Model Type: Sentence Transformer
Base model: TechWolf/JobBERT-v3
Maximum Sequence Length: 128 tokens
Output Dimensionality: 1024 dimensions
Similarity Function: Cosine Similarity
Supported Modality: Text

Model Sources

Documentation: Sentence Transformers Documentation
Repository: Sentence Transformers on GitHub
Hugging Face: Sentence Transformers on Hugging Face

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'transformer_task': 'feature-extraction', 'modality_config': {'text': {'method': 'forward', 'method_output_name': 'last_hidden_state'}}, 'module_output_name': 'token_embeddings', 'architecture': 'XLMRobertaModel'})
  (1): Pooling({'embedding_dimension': 768, 'pooling_mode': 'mean', 'include_prompt': True})
  (2): Router(
    default_route='anchor'
    (sub_modules): ModuleDict(
      (anchor): Sequential(
        (0): Dense({'in_features': 768, 'out_features': 1024, 'bias': True, 'activation_function': 'torch.nn.modules.activation.Tanh', 'module_input_name': 'sentence_embedding', 'module_output_name': 'sentence_embedding'})
      )
      (positive): Sequential(
        (0): Dense({'in_features': 768, 'out_features': 1024, 'bias': True, 'activation_function': 'torch.nn.modules.activation.Tanh', 'module_input_name': 'sentence_embedding', 'module_output_name': 'sentence_embedding'})
      )
    )
  )
)

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

# Download from the 🤗 Hub
model = SentenceTransformer("sentence_transformers_model_id")
# Run inference
queries = [
    'cybersecurity risk manager',
]
documents = [
    'cybersecurity risk assurance consultant',
    'konfigurátorka aplikácií',
    'vedúci predajne obuvi a koženej galantérie',
]
query_embeddings = model.encode_query(queries)
document_embeddings = model.encode_document(documents)
print(query_embeddings.shape, document_embeddings.shape)
# [1, 1024] [3, 1024]

# Get the similarity scores for the embeddings
similarities = model.similarity(query_embeddings, document_embeddings)
print(similarities.shape)
# [1, 3]

Training Details

Training Dataset

Unnamed Dataset

Size: 93,425 training samples
Columns: anchor and positive
Approximate statistics based on the first 1000 samples:
anchor positive
type string string
details
min: 3 tokens
mean: 8.03 tokens
max: 24 tokens

min: 4 tokens
mean: 30.21 tokens
max: 128 tokens

	anchor	positive
type	string	string
details	min: 3 tokens mean: 8.03 tokens max: 24 tokens	min: 4 tokens mean: 30.21 tokens max: 128 tokens

Samples:

anchor	positive
`laborant botanik`	`laborantka botanička`
`průvodčí vlaků v osobní dopravě`	`průvodčí osobní přepravy`
`vývojářka softwaru vestavěných systémů`	`vývojářka softwaru vestavěných zařízení`

Loss: MultipleNegativesRankingLoss with these parameters:

{
    "scale": 20.0,
    "similarity_fct": "cos_sim",
    "gather_across_devices": false,
    "directions": [
        "query_to_doc"
    ],
    "partition_mode": "joint",
    "hardness_mode": null,
    "hardness_strength": 0.0
}

Evaluation Dataset

Unnamed Dataset

Size: 8,686 evaluation samples
Columns: anchor and positive
Approximate statistics based on the first 1000 samples:
anchor positive
type string string
details
min: 3 tokens
mean: 8.01 tokens
max: 24 tokens

min: 4 tokens
mean: 32.34 tokens
max: 128 tokens

	anchor	positive
type	string	string
details	min: 3 tokens mean: 8.01 tokens max: 24 tokens	min: 4 tokens mean: 32.34 tokens max: 128 tokens

Samples:

anchor	positive
`instruktorka řízení osobních automobilů`	`instruktor řízení osobních automobilů`
`technička námořní mechatroniky`	`technolog námořní mechatroniky`
`specialista zahraničního obchodu v oblasti kancelářského nábytku`	`špecialistka v oblasti dovozu a vývozu kancelárskeho nábytku`

Loss: MultipleNegativesRankingLoss with these parameters:

{
    "scale": 20.0,
    "similarity_fct": "cos_sim",
    "gather_across_devices": false,
    "directions": [
        "query_to_doc"
    ],
    "partition_mode": "joint",
    "hardness_mode": null,
    "hardness_strength": 0.0
}

Training Hyperparameters

Non-Default Hyperparameters

per_device_train_batch_size: 128
learning_rate: 2e-05
warmup_steps: 0.1
fp16: True
per_device_eval_batch_size: 128
load_best_model_at_end: True
dataloader_drop_last: True
dataloader_num_workers: 4
router_mapping: {'anchor': 'anchor', 'positive': 'positive'}

All Hyperparameters

Click to expand

per_device_train_batch_size: 128
num_train_epochs: 3
max_steps: -1
learning_rate: 2e-05
lr_scheduler_type: linear
lr_scheduler_kwargs: None
warmup_steps: 0.1
optim: adamw_torch_fused
optim_args: None
weight_decay: 0.0
adam_beta1: 0.9
adam_beta2: 0.999
adam_epsilon: 1e-08
optim_target_modules: None
gradient_accumulation_steps: 1
average_tokens_across_devices: True
max_grad_norm: 1.0
label_smoothing_factor: 0.0
bf16: False
fp16: True
bf16_full_eval: False
fp16_full_eval: False
tf32: None
gradient_checkpointing: False
gradient_checkpointing_kwargs: None
torch_compile: False
torch_compile_backend: None
torch_compile_mode: None
use_liger_kernel: False
liger_kernel_config: None
use_cache: False
neftune_noise_alpha: None
torch_empty_cache_steps: None
auto_find_batch_size: False
log_on_each_node: True
logging_nan_inf_filter: True
include_num_input_tokens_seen: no
log_level: passive
log_level_replica: warning
disable_tqdm: False
project: huggingface
trackio_space_id: trackio
per_device_eval_batch_size: 128
prediction_loss_only: True
eval_on_start: False
eval_do_concat_batches: True
eval_use_gather_object: False
eval_accumulation_steps: None
include_for_metrics: []
batch_eval_metrics: False
save_only_model: False
save_on_each_node: False
enable_jit_checkpoint: False
push_to_hub: False
hub_private_repo: None
hub_model_id: None
hub_strategy: every_save
hub_always_push: False
hub_revision: None
load_best_model_at_end: True
ignore_data_skip: False
restore_callback_states_from_checkpoint: False
full_determinism: False
seed: 42
data_seed: None
use_cpu: False
accelerator_config: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
parallelism_config: None
dataloader_drop_last: True
dataloader_num_workers: 4
dataloader_pin_memory: True
dataloader_persistent_workers: False
dataloader_prefetch_factor: None
remove_unused_columns: True
label_names: None
train_sampling_strategy: random
length_column_name: length
ddp_find_unused_parameters: None
ddp_bucket_cap_mb: None
ddp_broadcast_buffers: False
ddp_backend: None
ddp_timeout: 1800
fsdp: []
fsdp_config: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
deepspeed: None
debug: []
skip_memory_metrics: True
do_predict: False
resume_from_checkpoint: None
warmup_ratio: None
local_rank: -1
prompts: None
batch_sampler: batch_sampler
multi_dataset_batch_sampler: proportional
router_mapping: {'anchor': 'anchor', 'positive': 'positive'}
learning_rate_mapping: {}

Training Logs

Epoch	Step	Training Loss	Validation Loss
0.0686	50	1.1902	-
0.1372	100	0.9478	-
0.2058	150	0.7830	-
0.2743	200	0.7300	-
0.3429	250	0.6671	0.9028
0.4115	300	0.6276	-
0.4801	350	0.5243	-
0.5487	400	0.5387	-
0.6173	450	0.5103	-
0.6859	500	0.4896	0.8111
0.7545	550	0.4634	-
0.8230	600	0.4549	-
0.8916	650	0.4426	-
0.9602	700	0.4225	-
1.0288	750	0.3886	0.8071
1.0974	800	0.3586	-
1.1660	850	0.3403	-
1.2346	900	0.3548	-
1.3032	950	0.3572	-
1.3717	1000	0.3491	0.7716
1.4403	1050	0.3446	-
1.5089	1100	0.3374	-
1.5775	1150	0.3297	-
1.6461	1200	0.2979	-
1.7147	1250	0.3281	0.7645
1.7833	1300	0.3010	-
1.8519	1350	0.3111	-
1.9204	1400	0.3072	-
1.9890	1450	0.3178	-
2.0576	1500	0.2969	0.7676
2.1262	1550	0.2701	-
2.1948	1600	0.2561	-
2.2634	1650	0.2661	-
2.3320	1700	0.2525	-
2.4005	1750	0.262	0.7628

The bold row denotes the saved checkpoint.

Training Time

Training: 8.8 minutes

Framework Versions

Python: 3.12.13
Sentence Transformers: 5.4.1
Transformers: 5.5.4
PyTorch: 2.11.0+cu130
Accelerate: 1.13.0
Datasets: 4.8.4
Tokenizers: 0.22.2

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{oord2019representationlearningcontrastivepredictive,
      title={Representation Learning with Contrastive Predictive Coding},
      author={Aaron van den Oord and Yazhe Li and Oriol Vinyals},
      year={2019},
      eprint={1807.03748},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/1807.03748},
}