Ehsanl/cncpt_e5

SentenceTransformer based on intfloat/multilingual-e5-large-instruct

This is a sentence-transformers model finetuned from intfloat/multilingual-e5-large-instruct on the json dataset. It maps inputs to a 1024-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, classification, clustering, and more.

Model Details

Model Description

• Model Type: Sentence Transformer
• Base model: intfloat/multilingual-e5-large-instruct
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 1024 dimensions
• Similarity Function: Cosine Similarity
• Supported Modality: Text
• Training Dataset:
  • json

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': 1024, 'pooling_mode': 'mean', 'include_prompt': True})
  (2): Normalize({'module_input_name': 'sentence_embedding', 'module_output_name': 'sentence_embedding'})
)

Evaluation

Metrics

Rank Zero IR

• Dataset: concept-dev
• Evaluated with train_retriever_bidirlm.RankZeroIREvaluator

Metric	Value
cosine_accuracy@1	0.223
cosine_accuracy@5	0.505
cosine_accuracy@10	0.615
cosine_precision@1	0.223
cosine_precision@10	0.0838
cosine_precision@25	0.0479
cosine_recall@1	0.0525
cosine_recall@10	0.1952
cosine_recall@25	0.2725
cosine_ndcg@10	0.1847
cosine_mrr@10	0.3442
cosine_map@100	0.1199

Training Details

Training Dataset

json

Training Hyperparameters

Non-Default Hyperparameters

• per_device_train_batch_size: 512
• num_train_epochs: 1.0
• learning_rate: 2e-05
• lr_scheduler_type: cosine
• warmup_steps: 58
• weight_decay: 0.01
• bf16: True
• tf32: True
• per_device_eval_batch_size: 512
• hub_private_repo: True
• hub_strategy: checkpoint
• load_best_model_at_end: True
• dataloader_num_workers: 2
• dataloader_pin_memory: False
• ddp_find_unused_parameters: False

All Hyperparameters

Click to expand

• per_device_train_batch_size: 512
• num_train_epochs: 1.0
• max_steps: -1
• learning_rate: 2e-05
• lr_scheduler_type: cosine
• lr_scheduler_kwargs: None
• warmup_steps: 58
• optim: adamw_torch_fused
• optim_args: None
• weight_decay: 0.01
• 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: True
• fp16: False
• bf16_full_eval: False
• fp16_full_eval: False
• tf32: True
• 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: None
• trackio_bucket_id: None
• trackio_static_space_id: None
• per_device_eval_batch_size: 512
• 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: True
• hub_model_id: None
• hub_strategy: checkpoint
• 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: False
• dataloader_num_workers: 2
• dataloader_pin_memory: False
• dataloader_persistent_workers: False
• dataloader_prefetch_factor: None
• dataloader_multiprocessing_context: None
• dataloader_in_order: True
• remove_unused_columns: True
• label_names: None
• train_sampling_strategy: random
• length_column_name: length
• ddp_find_unused_parameters: False
• ddp_bucket_cap_mb: None
• ddp_broadcast_buffers: False
• ddp_static_graph: None
• ddp_backend: None
• ddp_timeout: 1800
• fsdp: None
• fsdp_config: None
• deepspeed: None
• debug: []
• skip_memory_metrics: True
• do_predict: False
• resume_from_checkpoint: None
• local_rank: -1
• prompts: None
• batch_sampler: batch_sampler
• multi_dataset_batch_sampler: proportional
• router_mapping: {}
• learning_rate_mapping: {}
• warmup_ratio: None

Training Logs

Epoch	Step	Training Loss	concept-dev_cosine_ndcg@10
0.0017	1	8.2978	-
0.0173	10	8.1293	-
0.0347	20	7.7957	-
0.0520	30	7.3246	-
0.0693	40	6.9759	-
0.0867	50	6.7088	-
0.1040	60	6.5911	-
0.1213	70	6.4329	-
0.1386	80	6.3659	-
0.1560	90	6.2806	-
0.1733	100	6.2206	-
0.1906	110	6.1548	-
0.2080	120	6.1161	-
0.2253	130	6.0303	-
0.2426	140	6.0143	-
0.2600	150	5.9863	-
0.2773	160	5.9285	-
0.2946	170	5.9333	-
0.3120	180	5.8504	-
0.3293	190	5.8392	-
0.3466	200	5.7718	0.1746
0.3640	210	5.7754	-
0.3813	220	5.7743	-
0.3986	230	5.7048	-
0.4159	240	5.6923	-
0.4333	250	5.6670	-
0.4506	260	5.7028	-
0.4679	270	5.6281	-
0.4853	280	5.6590	-
0.5026	290	5.6350	-
0.5199	300	5.6218	-
0.5373	310	5.6122	-
0.5546	320	5.5454	-
0.5719	330	5.5093	-
0.5893	340	5.5001	-
0.6066	350	5.5197	-
0.6239	360	5.5142	-
0.6412	370	5.5230	-
0.6586	380	5.4936	-
0.6759	390	5.4790	-
0.6932	400	5.5171	0.1827
0.7106	410	5.4625	-
0.7279	420	5.4563	-
0.7452	430	5.4652	-
0.7626	440	5.4686	-
0.7799	450	5.4186	-
0.7972	460	5.4356	-
0.8146	470	5.4426	-
0.8319	480	5.4641	-
0.8492	490	5.4589	-
0.8666	500	5.4399	-
0.8839	510	5.4501	-
0.9012	520	5.4281	-
0.9185	530	5.4175	-
0.9359	540	5.4407	-
0.9532	550	5.4569	-
0.9705	560	5.4252	-
0.9879	570	5.4485	-
1.0	577	-	0.1847

• The bold row denotes the saved checkpoint.

Training Time

• Training: 1.3 hours
• Evaluation: 17.8 seconds
• Total: 1.3 hours

Framework Versions

• Python: 3.11.16
• Sentence Transformers: 6.1.0
• Transformers: 5.17.0
• PyTorch: 2.8.0+cu128
• Accelerate: 1.15.0
• Datasets: 5.0.1
• Tokenizers: 0.23.2

Additional Resources

• Training and Finetuning Embedding Models with Sentence Transformers: the end-to-end guide for training or finetuning Sentence Transformer models.
• Introduction to Matryoshka Embedding Models: variable-size embeddings that can be truncated with minimal quality loss.
• Binary and Scalar Embedding Quantization for Significantly Faster & Cheaper Retrieval: post-training compression of embedding vectors.
• Multimodal Embedding & Reranker Models with Sentence Transformers: use text, image, audio, and video models through the same API.
• Training and Finetuning Multimodal Embedding & Reranker Models with Sentence Transformers: train multimodal embedding models, with a Visual Document Retrieval walkthrough.

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",
}

WeightedMultiPositiveCachedLoss

@misc{gao2021scaling,
    title={Scaling Deep Contrastive Learning Batch Size under Memory Limited Setup},
    author={Luyu Gao and Yunyi Zhang and Jiawei Han and Jamie Callan},
    year={2021},
    eprint={2101.06983},
    archivePrefix={arXiv},
    primaryClass={cs.LG}
}

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},
}