modernbert-embed-base-bible

This is a sentence-transformers model finetuned from nomic-ai/modernbert-embed-base on the json dataset. 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: nomic-ai/modernbert-embed-base
• Maximum Sequence Length: 8192 tokens
• Output Dimensionality: 768 dimensions
• Similarity Function: Cosine Similarity
• Training Dataset:
  • json
• Language: fr
• License: apache-2.0

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': 8192, 'do_lower_case': False}) with Transformer model: ModernBertModel 
  (1): Pooling({'word_embedding_dimension': 768, '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()
)

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("Steve77/modernbert-embed-base-bible")
# Run inference
sentences = [
    "Quelles tâches les Lévites devaient-ils accomplir dans le service de la maison de l'Éternel?",
    "Ils devaient prendre soin des parvis et des chambres, purifier toutes les choses saintes, s'occuper des pains de proposition, de la fleur de farine pour les offrandes, des galettes sans levain, des gâteaux cuits sur la plaque et des gâteaux frits, et de toutes les mesures de capacité et de longueur.",
    "Les chefs des maisons paternelles, les chefs des tribus d'Israël, les chefs de milliers et de centaines, et les intendants du roi.",
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]

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

Evaluation

Metrics

Information Retrieval

• Datasets: dim_768, dim_512, dim_256, dim_128 and dim_64
• Evaluated with InformationRetrievalEvaluator

Metric	dim_768	dim_512	dim_256	dim_128	dim_64
cosine_accuracy@1	0.175	0.1716	0.1602	0.1437	0.1107
cosine_accuracy@3	0.2484	0.2402	0.2276	0.2047	0.1595
cosine_accuracy@5	0.2762	0.272	0.2603	0.2331	0.1862
cosine_accuracy@10	0.3203	0.3162	0.3075	0.2794	0.2272
cosine_precision@1	0.175	0.1716	0.1602	0.1437	0.1107
cosine_precision@3	0.0828	0.0801	0.0759	0.0682	0.0532
cosine_precision@5	0.0552	0.0544	0.0521	0.0466	0.0372
cosine_precision@10	0.032	0.0316	0.0308	0.0279	0.0227
cosine_recall@1	0.175	0.1716	0.1602	0.1437	0.1107
cosine_recall@3	0.2484	0.2402	0.2276	0.2047	0.1595
cosine_recall@5	0.2762	0.272	0.2603	0.2331	0.1862
cosine_recall@10	0.3203	0.3162	0.3075	0.2794	0.2272
cosine_ndcg@10	0.2443	0.2395	0.2284	0.2057	0.1633
cosine_ndcg@15	0.2525	0.2464	0.2357	0.2141	0.17
cosine_ndcg@20	0.2574	0.2517	0.2405	0.2194	0.1746
cosine_mrr@10	0.2205	0.2155	0.2038	0.1829	0.1435
cosine_map@100	0.2276	0.2226	0.2108	0.1901	0.15

Training Details

Training Dataset

json

• Dataset: json
• Size: 47,560 training samples
• Columns: anchor and positive
• Approximate statistics based on the first 1000 samples:

  	anchor	positive
  type	string	string
  details	min: 8 tokens mean: 21.11 tokens max: 45 tokens	min: 3 tokens mean: 24.84 tokens max: 108 tokens

• Samples:

  anchor	positive
  Quels sont les noms des fils de Schobal?	Aljan, Manahath, Ébal, Schephi et Onam
  Quels sont les noms des fils de Tsibeon?	Ajja et Ana
  Qui est le fils d'Ana?	Dischon

• Loss: MatryoshkaLoss with these parameters:

  {
      "loss": "MultipleNegativesRankingLoss",
      "matryoshka_dims": [
          768,
          512,
          256,
          128,
          64
      ],
      "matryoshka_weights": [
          1,
          1,
          1,
          1,
          1
      ],
      "n_dims_per_step": -1
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: epoch
• per_device_train_batch_size: 16
• per_device_eval_batch_size: 16
• gradient_accumulation_steps: 16
• learning_rate: 2e-05
• num_train_epochs: 4
• lr_scheduler_type: cosine
• warmup_ratio: 0.1
• bf16: True
• load_best_model_at_end: True
• optim: adamw_torch_fused
• batch_sampler: no_duplicates

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: epoch
• prediction_loss_only: True
• per_device_train_batch_size: 16
• per_device_eval_batch_size: 16
• per_gpu_train_batch_size: None
• per_gpu_eval_batch_size: None
• gradient_accumulation_steps: 16
• 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: 4
• max_steps: -1
• lr_scheduler_type: cosine
• 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_fused
• 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: False
• resume_from_checkpoint: None
• hub_model_id: None
• hub_strategy: every_save
• hub_private_repo: None
• 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	dim_768_cosine_ndcg@20	dim_512_cosine_ndcg@20	dim_256_cosine_ndcg@20	dim_128_cosine_ndcg@20	dim_64_cosine_ndcg@20
0.0538	10	12.274	-	-	-	-	-
0.1076	20	11.5084	-	-	-	-	-
0.1615	30	10.5276	-	-	-	-	-
0.2153	40	9.0432	-	-	-	-	-
0.2691	50	7.572	-	-	-	-	-
0.3229	60	7.7696	-	-	-	-	-
0.3767	70	6.5673	-	-	-	-	-
0.4305	80	6.6586	-	-	-	-	-
0.4844	90	5.5276	-	-	-	-	-
0.5382	100	5.9891	-	-	-	-	-
0.5920	110	5.2983	-	-	-	-	-
0.6458	120	5.6242	-	-	-	-	-
0.6996	130	5.498	-	-	-	-	-
0.7534	140	4.4201	-	-	-	-	-
0.8073	150	4.3818	-	-	-	-	-
0.8611	160	4.2175	-	-	-	-	-
0.9149	170	4.2341	-	-	-	-	-
0.9687	180	4.3349	-	-	-	-	-
0.9956	185	-	0.2664	0.2607	0.2508	0.2263	0.1796
1.0269	190	4.6803	-	-	-	-	-
1.0807	200	3.877	-	-	-	-	-
1.1345	210	4.0309	-	-	-	-	-
1.1884	220	4.0755	-	-	-	-	-
1.2422	230	3.9068	-	-	-	-	-
1.2960	240	4.188	-	-	-	-	-
1.3498	250	4.3417	-	-	-	-	-
1.4036	260	4.0526	-	-	-	-	-
1.4575	270	3.3933	-	-	-	-	-
1.5113	280	3.8309	-	-	-	-	-
1.5651	290	3.5633	-	-	-	-	-
1.6189	300	3.8179	-	-	-	-	-
1.6727	310	4.0671	-	-	-	-	-
1.7265	320	3.3919	-	-	-	-	-
1.7804	330	2.6578	-	-	-	-	-
1.8342	340	2.6953	-	-	-	-	-
1.8880	350	2.8858	-	-	-	-	-
1.9418	360	2.8933	-	-	-	-	-
1.9956	370	2.9603	0.2775	0.2737	0.2637	0.2402	0.1916
2.0538	380	3.3361	-	-	-	-	-
2.1076	390	2.7904	-	-	-	-	-
2.1615	400	3.0108	-	-	-	-	-
2.2153	410	2.8917	-	-	-	-	-
2.2691	420	3.0295	-	-	-	-	-
2.3229	430	3.5609	-	-	-	-	-
2.3767	440	2.7722	-	-	-	-	-
2.4305	450	3.2115	-	-	-	-	-
2.4844	460	2.6333	-	-	-	-	-
2.5382	470	3.2503	-	-	-	-	-
2.5920	480	2.7708	-	-	-	-	-
2.6458	490	3.167	-	-	-	-	-
2.6996	500	3.1447	-	-	-	-	-
2.7534	510	2.0428	-	-	-	-	-
2.8073	520	2.0001	-	-	-	-	-
2.8611	530	2.0826	-	-	-	-	-
2.9149	540	2.0853	-	-	-	-	-
2.9687	550	2.2365	-	-	-	-	-
2.9956	555	-	0.2660	0.2604	0.2509	0.2266	0.1810
3.0269	560	2.762	-	-	-	-	-
3.0807	570	2.1219	-	-	-	-	-
3.1345	580	2.2908	-	-	-	-	-
3.1884	590	2.6195	-	-	-	-	-
3.2422	600	2.3468	-	-	-	-	-
3.2960	610	2.7504	-	-	-	-	-
3.3498	620	2.9486	-	-	-	-	-
3.4036	630	2.7281	-	-	-	-	-
3.4575	640	2.188	-	-	-	-	-
3.5113	650	2.5494	-	-	-	-	-
3.5651	660	2.426	-	-	-	-	-
3.6189	670	2.6478	-	-	-	-	-
3.6727	680	2.9209	-	-	-	-	-
3.7265	690	2.3512	-	-	-	-	-
3.7804	700	1.6746	-	-	-	-	-
3.8342	710	1.739	-	-	-	-	-
3.8880	720	1.951	-	-	-	-	-
3.9418	730	1.9886	-	-	-	-	-
3.9956	740	2.1022	0.2574	0.2517	0.2405	0.2194	0.1746

• The bold row denotes the saved checkpoint.

Environmental Impact

Carbon emissions were measured using CodeCarbon.

• Energy Consumed: 0.205 kWh
• Carbon Emitted: 0.011 kg of CO2
• Hours Used: 6.806 hours

Training Hardware

• On Cloud: No
• GPU Model: 1 x NVIDIA GeForce GTX 1660 Ti
• CPU Model: Intel(R) Core(TM) i7-9750H CPU @ 2.60GHz
• RAM Size: 7.68 GB

Framework Versions

• Python: 3.11.11
• Sentence Transformers: 3.3.1
• Transformers: 4.48.0.dev0
• PyTorch: 2.5.1
• Accelerate: 1.2.1
• Datasets: 2.19.1
• 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",
}

MatryoshkaLoss

@misc{kusupati2024matryoshka,
    title={Matryoshka Representation Learning},
    author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi},
    year={2024},
    eprint={2205.13147},
    archivePrefix={arXiv},
    primaryClass={cs.LG}
}

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}
}