pj-mathematician/JobGTE-7b-Lora

Job - Job matching finetuned Alibaba-NLP/gte-Qwen2-7B-instruct

Best performing model on TalentCLEF 2025 Task A. Use it for multilingual job title matching

Model Details

Model Description

• Model Type: Sentence Transformer
• Base model: Alibaba-NLP/gte-Qwen2-7B-instruct
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 3584 dimensions
• Similarity Function: Cosine Similarity
• Training Datasets:
  • full_en
  • full_de
  • full_es
  • full_zh
  • mix

Model Sources

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

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: Qwen2Model 
  (1): Pooling({'word_embedding_dimension': 3584, 'pooling_mode_cls_token': False, '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': True, '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

# Download from the 🤗 Hub
model = SentenceTransformer("pj-mathematician/JobGTE-7b-Lora")
# Run inference
sentences = [
    'Volksvertreter',
    'Parlamentarier',
    'Oberbürgermeister',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 3584]

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

Training Details

Training Datasets

full_en

full_en

• Dataset: full_en
• Size: 28,880 training samples
• Columns: anchor and positive
• Approximate statistics based on the first 1000 samples:

  	anchor	positive
  type	string	string
  details	min: 2 tokens mean: 4.4 tokens max: 9 tokens	min: 2 tokens mean: 4.42 tokens max: 10 tokens

• Samples:

  anchor	positive
  air commodore	flight lieutenant
  command and control officer	flight officer
  air commodore	command and control officer

• Loss: CachedGISTEmbedLoss with these parameters:

  {'guide': SentenceTransformer(
    (0): Transformer({'max_seq_length': 128, 'do_lower_case': False}) with Transformer model: BertModel 
    (1): Pooling({'word_embedding_dimension': 384, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, '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()
  ), 'temperature': 0.01, 'mini_batch_size': 64, 'margin_strategy': 'absolute', 'margin': 0.0}
  

full_de

full_de

• Dataset: full_de
• Size: 23,023 training samples
• Columns: anchor and positive
• Approximate statistics based on the first 1000 samples:

  	anchor	positive
  type	string	string
  details	min: 2 tokens mean: 9.11 tokens max: 33 tokens	min: 2 tokens mean: 9.41 tokens max: 33 tokens

• Samples:

  anchor	positive
  Staffelkommandantin	Kommodore
  Luftwaffenoffizierin	Luftwaffenoffizier/Luftwaffenoffizierin
  Staffelkommandantin	Luftwaffenoffizierin

• Loss: CachedGISTEmbedLoss with these parameters:

  {'guide': SentenceTransformer(
    (0): Transformer({'max_seq_length': 128, 'do_lower_case': False}) with Transformer model: BertModel 
    (1): Pooling({'word_embedding_dimension': 384, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, '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()
  ), 'temperature': 0.01, 'mini_batch_size': 64, 'margin_strategy': 'absolute', 'margin': 0.0}
  

full_es

full_es

• Dataset: full_es
• Size: 20,724 training samples
• Columns: anchor and positive
• Approximate statistics based on the first 1000 samples:

  	anchor	positive
  type	string	string
  details	min: 3 tokens mean: 9.42 tokens max: 35 tokens	min: 3 tokens mean: 9.18 tokens max: 35 tokens

• Samples:

  anchor	positive
  jefe de escuadrón	instructor
  comandante de aeronave	instructor de simulador
  instructor	oficial del Ejército del Aire

• Loss: CachedGISTEmbedLoss with these parameters:

  {'guide': SentenceTransformer(
    (0): Transformer({'max_seq_length': 128, 'do_lower_case': False}) with Transformer model: BertModel 
    (1): Pooling({'word_embedding_dimension': 384, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, '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()
  ), 'temperature': 0.01, 'mini_batch_size': 64, 'margin_strategy': 'absolute', 'margin': 0.0}
  

full_zh

full_zh

• Dataset: full_zh
• Size: 30,401 training samples
• Columns: anchor and positive
• Approximate statistics based on the first 1000 samples:

  	anchor	positive
  type	string	string
  details	min: 3 tokens mean: 4.7 tokens max: 12 tokens	min: 3 tokens mean: 5.04 tokens max: 19 tokens

• Samples:

  anchor	positive
  技术总监	技术和运营总监
  技术总监	技术主管
  技术总监	技术艺术总监

• Loss: CachedGISTEmbedLoss with these parameters:

  {'guide': SentenceTransformer(
    (0): Transformer({'max_seq_length': 128, 'do_lower_case': False}) with Transformer model: BertModel 
    (1): Pooling({'word_embedding_dimension': 384, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, '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()
  ), 'temperature': 0.01, 'mini_batch_size': 64, 'margin_strategy': 'absolute', 'margin': 0.0}
  

mix

mix

• Dataset: mix
• Size: 21,760 training samples
• Columns: anchor and positive
• Approximate statistics based on the first 1000 samples:

  	anchor	positive
  type	string	string
  details	min: 1 tokens mean: 4.98 tokens max: 14 tokens	min: 1 tokens mean: 7.22 tokens max: 27 tokens

• Samples:

  anchor	positive
  technical manager	Technischer Direktor für Bühne, Film und Fernsehen
  head of technical	directora técnica
  head of technical department	技术艺术总监

• Loss: CachedGISTEmbedLoss with these parameters:

  {'guide': SentenceTransformer(
    (0): Transformer({'max_seq_length': 128, 'do_lower_case': False}) with Transformer model: BertModel 
    (1): Pooling({'word_embedding_dimension': 384, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, '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()
  ), 'temperature': 0.01, 'mini_batch_size': 64, 'margin_strategy': 'absolute', 'margin': 0.0}
  

Training Hyperparameters

Non-Default Hyperparameters

• per_device_train_batch_size: 128
• per_device_eval_batch_size: 128
• gradient_accumulation_steps: 2
• num_train_epochs: 2
• warmup_ratio: 0.05
• log_on_each_node: False
• fp16: True
• dataloader_num_workers: 4
• fsdp: ['full_shard', 'auto_wrap']
• fsdp_config: {'transformer_layer_cls_to_wrap': ['Qwen2DecoderLayer'], 'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
• ddp_find_unused_parameters: True
• gradient_checkpointing: True
• batch_sampler: no_duplicates

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: no
• prediction_loss_only: True
• per_device_train_batch_size: 128
• per_device_eval_batch_size: 128
• per_gpu_train_batch_size: None
• per_gpu_eval_batch_size: None
• gradient_accumulation_steps: 2
• eval_accumulation_steps: None
• torch_empty_cache_steps: None
• learning_rate: 5e-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: 2
• max_steps: -1
• lr_scheduler_type: linear
• lr_scheduler_kwargs: {}
• warmup_ratio: 0.05
• warmup_steps: 0
• log_level: passive
• log_level_replica: warning
• log_on_each_node: False
• 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: False
• fp16: True
• 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: True
• dataloader_num_workers: 4
• dataloader_prefetch_factor: None
• past_index: -1
• disable_tqdm: False
• remove_unused_columns: True
• label_names: None
• load_best_model_at_end: False
• ignore_data_skip: False
• fsdp: ['full_shard', 'auto_wrap']
• fsdp_min_num_params: 0
• fsdp_config: {'transformer_layer_cls_to_wrap': ['Qwen2DecoderLayer'], 'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
• tp_size: 0
• 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: True
• 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: False
• resume_from_checkpoint: None
• hub_model_id: None
• hub_strategy: every_save
• hub_private_repo: None
• hub_always_push: False
• gradient_checkpointing: True
• 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
• 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
0.0165	1	4.5178
0.0331	2	3.8803
0.0496	3	2.8882
0.0661	4	4.5362
0.0826	5	3.6406
0.0992	6	3.5285
0.1157	7	4.1398
0.1322	8	4.1543
0.1488	9	4.4487
0.1653	10	4.7408
0.1818	11	2.1874
0.1983	12	3.3176
0.2149	13	2.8286
0.2314	14	2.87
0.2479	15	2.4834
0.2645	16	2.7856
0.2810	17	3.1948
0.2975	18	2.1755
0.3140	19	1.9861
0.3306	20	2.0536
0.3471	21	2.7626
0.3636	22	1.6489
0.3802	23	2.078
0.3967	24	1.5864
0.4132	25	1.8815
0.4298	26	1.8041
0.4463	27	1.7482
0.4628	28	1.191
0.4793	29	1.4166
0.4959	30	1.3215
0.5124	31	1.2907
0.5289	32	1.1294
0.5455	33	1.1586
0.5620	34	1.551
0.5785	35	1.3628
0.5950	36	0.9899
0.6116	37	1.1846
0.6281	38	1.2721
0.6446	39	1.1261
0.6612	40	0.9535
0.6777	41	1.2086
0.6942	42	0.7472
0.7107	43	1.0324
0.7273	44	1.0397
0.7438	45	1.185
0.7603	46	1.2112
0.7769	47	0.84
0.7934	48	0.9286
0.8099	49	0.8689
0.8264	50	0.9546
0.8430	51	0.8283
0.8595	52	0.757
0.8760	53	0.9199
0.8926	54	0.7404
0.9091	55	1.0995
0.9256	56	0.8231
0.9421	57	0.6297
0.9587	58	0.9869
0.9752	59	0.9597
0.9917	60	0.7025
1.0	61	0.4866

Framework Versions

• Python: 3.11.11
• Sentence Transformers: 4.1.0
• Transformers: 4.51.3
• PyTorch: 2.6.0+cu124
• Accelerate: 1.6.0
• Datasets: 3.5.0
• Tokenizers: 0.21.1

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",
}