juanlofer/bge-base-fastdds-summaries-20epochs-666seed

BGE base Fast-DDS summaries

This is a sentence-transformers model finetuned from BAAI/bge-base-en-v1.5. 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: BAAI/bge-base-en-v1.5
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 768 tokens
• Similarity Function: Cosine Similarity
• Language: en
• 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': 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

# Download from the 🤗 Hub
model = SentenceTransformer("juanlofer/bge-base-fastdds-summaries-20epochs-666seed")
# Run inference
sentences = [
    'The transport layer provides communication services between DDS entities, using UDPv4, UDPv6, TCPv4, TCPv6, and SHM transports.',
    '* **TCPv4**: TCP communication over IPv4 (see TCP Transport).',
    'The following table shows the supported primitive types and their\ncorresponding "TypeKind". The "TypeKind" is used to query the\nDynamicTypeBuilderFactory for the specific primitive DynamicType.',
]
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

• Dataset: dim_768
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.3341
cosine_accuracy@3	0.4455
cosine_accuracy@5	0.5035
cosine_accuracy@10	0.5661
cosine_precision@1	0.3341
cosine_precision@3	0.1485
cosine_precision@5	0.1007
cosine_precision@10	0.0566
cosine_recall@1	0.3341
cosine_recall@3	0.4455
cosine_recall@5	0.5035
cosine_recall@10	0.5661
cosine_ndcg@10	0.4437
cosine_mrr@10	0.4054
cosine_map@100	0.416

Information Retrieval

• Dataset: dim_512
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.3364
cosine_accuracy@3	0.4478
cosine_accuracy@5	0.4965
cosine_accuracy@10	0.5777
cosine_precision@1	0.3364
cosine_precision@3	0.1493
cosine_precision@5	0.0993
cosine_precision@10	0.0578
cosine_recall@1	0.3364
cosine_recall@3	0.4478
cosine_recall@5	0.4965
cosine_recall@10	0.5777
cosine_ndcg@10	0.4463
cosine_mrr@10	0.4057
cosine_map@100	0.4154

Information Retrieval

• Dataset: dim_256
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.3271
cosine_accuracy@3	0.4478
cosine_accuracy@5	0.4988
cosine_accuracy@10	0.5754
cosine_precision@1	0.3271
cosine_precision@3	0.1493
cosine_precision@5	0.0998
cosine_precision@10	0.0575
cosine_recall@1	0.3271
cosine_recall@3	0.4478
cosine_recall@5	0.4988
cosine_recall@10	0.5754
cosine_ndcg@10	0.4414
cosine_mrr@10	0.3997
cosine_map@100	0.4105

Information Retrieval

• Dataset: dim_128
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.3155
cosine_accuracy@3	0.4292
cosine_accuracy@5	0.4803
cosine_accuracy@10	0.5754
cosine_precision@1	0.3155
cosine_precision@3	0.1431
cosine_precision@5	0.0961
cosine_precision@10	0.0575
cosine_recall@1	0.3155
cosine_recall@3	0.4292
cosine_recall@5	0.4803
cosine_recall@10	0.5754
cosine_ndcg@10	0.4328
cosine_mrr@10	0.389
cosine_map@100	0.3994

Information Retrieval

• Dataset: dim_64
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.2854
cosine_accuracy@3	0.4153
cosine_accuracy@5	0.4687
cosine_accuracy@10	0.5568
cosine_precision@1	0.2854
cosine_precision@3	0.1384
cosine_precision@5	0.0937
cosine_precision@10	0.0557
cosine_recall@1	0.2854
cosine_recall@3	0.4153
cosine_recall@5	0.4687
cosine_recall@10	0.5568
cosine_ndcg@10	0.4098
cosine_mrr@10	0.3641
cosine_map@100	0.3744

Training Details

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: 20
• lr_scheduler_type: cosine
• warmup_ratio: 0.1
• fp16: True
• tf32: False
• 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
• 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: 20
• 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: False
• fp16: True
• fp16_opt_level: O1
• half_precision_backend: auto
• bf16_full_eval: False
• fp16_full_eval: False
• tf32: False
• 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: False
• hub_always_push: False
• gradient_checkpointing: False
• gradient_checkpointing_kwargs: None
• include_inputs_for_metrics: False
• 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
• batch_sampler: no_duplicates
• multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	dim_128_cosine_map@100	dim_256_cosine_map@100	dim_512_cosine_map@100	dim_64_cosine_map@100	dim_768_cosine_map@100
0.6584	10	5.9441	-	-	-	-	-
0.9877	15	-	0.3686	0.3792	0.3819	0.3414	0.3795
1.3128	20	4.7953	-	-	-	-	-
1.9712	30	3.77	0.3854	0.3963	0.3962	0.3682	0.3995
2.6255	40	2.9211	-	-	-	-	-
2.9547	45	-	0.3866	0.3919	0.3958	0.3759	0.3963
3.2798	50	2.4548	-	-	-	-	-
3.9383	60	2.0513	-	-	-	-	-
4.0041	61	-	0.3808	0.4018	0.3980	0.3647	0.3962
4.5926	70	1.5898	-	-	-	-	-
4.9877	76	-	0.3829	0.4029	0.4035	0.3625	0.4014
5.2469	80	1.4677	-	-	-	-	-
5.9053	90	1.1974	-	-	-	-	-
5.9712	91	-	0.3918	0.4006	0.4041	0.3654	0.4033
6.5597	100	0.9285	-	-	-	-	-
6.9547	106	-	0.3914	0.4019	0.4033	0.3678	0.4014
7.2140	110	0.9214	-	-	-	-	-
7.8724	120	0.8141	-	-	-	-	-
8.0041	122	-	0.3914	0.3993	0.4071	0.3670	0.4027
8.5267	130	0.6706	-	-	-	-	-
8.9877	137	-	0.3903	0.4033	0.4060	0.3721	0.4060
9.1811	140	0.6388	-	-	-	-	-
9.8395	150	0.5466	-	-	-	-	-
9.9712	152	-	0.3915	0.4020	0.4079	0.3673	0.4046
10.4938	160	0.466	-	-	-	-	-
10.9547	167	-	0.3963	0.4069	0.4112	0.3697	0.4078
11.1481	170	0.4709	-	-	-	-	-
11.8066	180	0.437	-	-	-	-	-
12.0041	183	-	0.4003	0.4051	0.4096	0.3701	0.4059
12.4609	190	0.3678	-	-	-	-	-
12.9877	198	-	0.3976	0.4075	0.4088	0.3713	0.4080
13.1152	200	0.3944	-	-	-	-	-
13.7737	210	0.361	-	-	-	-	-
13.9712	213	-	0.3966	0.4091	0.4096	0.3724	0.4107
14.4280	220	0.2977	-	-	-	-	-
14.9547	228	-	0.3979	0.4102	0.4149	0.3744	0.4143
15.0823	230	0.3306	-	-	-	-	-
15.7407	240	0.3075	-	-	-	-	-
16.0041	244	-	0.3991	0.4102	0.4156	0.3726	0.4148
16.3951	250	0.2777	-	-	-	-	-
16.9877	259	-	0.3990	0.4101	0.4154	0.3743	0.4167
17.0494	260	0.3044	-	-	-	-	-
17.7078	270	0.2885	-	-	-	-	-
17.9712	274	-	0.3991	0.4099	0.4153	0.3746	0.4167
18.3621	280	0.2862	-	-	-	-	-
18.9547	289	-	0.3994	0.4105	0.4154	0.3743	0.4156
19.0165	290	0.2974	-	-	-	-	-
19.6749	300	0.2648	0.3994	0.4105	0.4154	0.3744	0.4160

• The bold row denotes the saved checkpoint.

Framework Versions

• Python: 3.10.13
• Sentence Transformers: 3.0.1
• Transformers: 4.41.2
• PyTorch: 2.1.2
• Accelerate: 0.30.1
• Datasets: 2.19.1
• Tokenizers: 0.19.1

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",
}

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}
}