yyzheng00/snomed_triplet_800k

SentenceTransformer based on sentence-transformers/all-MiniLM-L6-v2

This is a sentence-transformers model finetuned from sentence-transformers/all-MiniLM-L6-v2 on the parquet dataset. It maps sentences & paragraphs to a 384-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: sentence-transformers/all-MiniLM-L6-v2
• Maximum Sequence Length: 256 tokens
• Output Dimensionality: 384 dimensions
• Similarity Function: Cosine Similarity
• Training Dataset:
  • parquet

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': 256, '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()
)

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("yyzheng00/snomed_triplet_800k")
# Run inference
sentences = [
    '|Adverse reaction caused by drug| : { |Causative agent| = |Digestant| }',
    'Adverse reaction caused by digestant (disorder)',
    'Ureteroscopic division of stricture of ureter (procedure)',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 384]

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

Evaluation

Metrics

Triplet

• Dataset: snomed_triplet_800k_3_4_3-dev
• Evaluated with TripletEvaluator

Metric	Value
cosine_accuracy	0.9983

Triplet

• Dataset: snomed_triplet_800k_3_4_3-dev
• Evaluated with TripletEvaluator

Metric	Value
cosine_accuracy	0.9984

Training Details

Training Dataset

parquet

• Dataset: parquet
• Size: 800,000 training samples
• Columns: anchor, positive, and negative
• Approximate statistics based on the first 1000 samples:

  	anchor	positive	negative
  type	string	string	string
  details	min: 19 tokens mean: 59.07 tokens max: 256 tokens	min: 4 tokens mean: 12.65 tokens max: 44 tokens	min: 4 tokens mean: 12.57 tokens max: 51 tokens

• Samples:

  anchor	positive	negative
  	Product containing lercanidipine	+
  	Product containing trovafloxacin	+
  	Product containing carboxylic acid and/or carboxylic acid derivative	+

• Loss: TripletLoss with these parameters:

  {
      "distance_metric": "TripletDistanceMetric.COSINE",
      "triplet_margin": 0.2
  }
  

Evaluation Dataset

parquet

• Dataset: parquet
• Size: 800,000 evaluation samples
• Columns: anchor, positive, and negative
• Approximate statistics based on the first 1000 samples:

  	anchor	positive	negative
  type	string	string	string
  details	min: 19 tokens mean: 58.9 tokens max: 253 tokens	min: 4 tokens mean: 12.41 tokens max: 49 tokens	min: 3 tokens mean: 12.43 tokens max: 38 tokens

• Samples:

  anchor	positive	negative
  	Hodgkin lymphoma, nodular lymphocyte predominance	+
  	Product containing bexarotene	+
  	Disease caused by Haemogregarinidae	: {

• Loss: TripletLoss with these parameters:

  {
      "distance_metric": "TripletDistanceMetric.COSINE",
      "triplet_margin": 0.2
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 16
• per_device_eval_batch_size: 16
• num_train_epochs: 1
• warmup_ratio: 0.1
• fp16: True
• batch_sampler: no_duplicates

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: steps
• 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: 1
• 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: 1
• max_steps: -1
• lr_scheduler_type: linear
• 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: 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: False
• 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
• 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

Click to expand

Epoch	Step	Training Loss	Validation Loss	snomed_triplet_800k_3_4_3-dev_cosine_accuracy
0.0033	100	0.0207	0.0115	0.9820
0.0067	200	0.0121	0.0093	0.9852
0.01	300	0.013	0.0082	0.9867
0.0133	400	0.0086	0.0076	0.9879
0.0167	500	0.0078	0.0072	0.9886
0.02	600	0.0084	0.0068	0.9893
0.0233	700	0.0066	0.0064	0.9901
0.0267	800	0.0077	0.0059	0.9905
0.03	900	0.0061	0.0059	0.9903
0.0333	1000	0.0069	0.0059	0.9906
0.0367	1100	0.0063	0.0055	0.9911
0.04	1200	0.0054	0.0055	0.9913
0.0433	1300	0.0062	0.0054	0.9917
0.0467	1400	0.0055	0.0053	0.9915
0.05	1500	0.0064	0.0051	0.9926
0.0533	1600	0.0062	0.0050	0.9928
0.0567	1700	0.0055	0.0047	0.9932
0.06	1800	0.0048	0.0052	0.9918
0.0633	1900	0.0056	0.0050	0.9929
0.0667	2000	0.006	0.0051	0.9926
0.07	2100	0.0061	0.0046	0.9932
0.0733	2200	0.0075	0.0045	0.9936
0.0767	2300	0.0055	0.0049	0.9923
0.08	2400	0.0043	0.0046	0.9935
0.0833	2500	0.006	0.0049	0.9923
0.0867	2600	0.0052	0.0048	0.9928
0.09	2700	0.006	0.0047	0.9927
0.0933	2800	0.0062	0.0042	0.9938
0.0967	2900	0.0056	0.0043	0.9942
0.1	3000	0.0049	0.0046	0.9929
0.1033	3100	0.0037	0.0043	0.9935
0.1067	3200	0.0061	0.0045	0.9929
0.11	3300	0.0045	0.0043	0.9934
0.1133	3400	0.0047	0.0047	0.9925
0.1167	3500	0.0068	0.0043	0.9938
0.12	3600	0.0054	0.0041	0.9935
0.1233	3700	0.0047	0.0041	0.9935
0.1267	3800	0.0062	0.0041	0.9938
0.13	3900	0.0029	0.0043	0.9935
0.1333	4000	0.0045	0.0039	0.9939
0.1367	4100	0.0046	0.0039	0.9943
0.14	4200	0.0048	0.0046	0.9928
0.1433	4300	0.0045	0.0043	0.9940
0.1467	4400	0.0058	0.0043	0.9935
0.15	4500	0.0045	0.0043	0.9930
0.1533	4600	0.0046	0.0040	0.9940
0.1567	4700	0.0049	0.0039	0.9944
0.16	4800	0.0047	0.0039	0.9940
0.1633	4900	0.0053	0.0047	0.9938
0.1667	5000	0.0042	0.0041	0.9934
0.17	5100	0.004	0.0039	0.9939
0.1733	5200	0.0037	0.0036	0.9944
0.1767	5300	0.005	0.0037	0.9942
0.18	5400	0.005	0.0035	0.9947
0.1833	5500	0.0047	0.0037	0.9944
0.1867	5600	0.0047	0.0035	0.9947
0.19	5700	0.0046	0.0037	0.9942
0.1933	5800	0.005	0.0036	0.9944
0.1967	5900	0.0052	0.0037	0.9945
0.2	6000	0.0044	0.0035	0.9952
0.2033	6100	0.0043	0.0034	0.9952
0.2067	6200	0.0046	0.0035	0.9947
0.21	6300	0.0059	0.0034	0.9948
0.2133	6400	0.0051	0.0035	0.9948
0.2167	6500	0.0032	0.0032	0.9950
0.22	6600	0.0031	0.0032	0.9951
0.2233	6700	0.003	0.0033	0.9951
0.2267	6800	0.0048	0.0034	0.9950
0.23	6900	0.0028	0.0038	0.9940
0.2333	7000	0.0035	0.0032	0.9951
0.2367	7100	0.0032	0.0032	0.9956
0.24	7200	0.0039	0.0032	0.9952
0.2433	7300	0.0046	0.0032	0.9951
0.2467	7400	0.0042	0.0033	0.9949
0.25	7500	0.0041	0.0032	0.9955
0.2533	7600	0.0043	0.0032	0.9954
0.2567	7700	0.0054	0.0031	0.9957
0.26	7800	0.0037	0.0033	0.9952
0.2633	7900	0.0042	0.0035	0.9948
0.2667	8000	0.0038	0.0031	0.9956
0.27	8100	0.0036	0.0031	0.9954
0.2733	8200	0.0037	0.0031	0.9956
0.2767	8300	0.0043	0.0030	0.9952
0.28	8400	0.004	0.0030	0.9957
0.2833	8500	0.0024	0.0030	0.9955
0.2867	8600	0.0032	0.0028	0.9959
0.29	8700	0.0045	0.0029	0.9957
0.2933	8800	0.0038	0.0030	0.9955
0.2967	8900	0.0057	0.0027	0.9960
0.3	9000	0.0043	0.0029	0.9957
0.3033	9100	0.0041	0.0029	0.9960
0.3067	9200	0.0028	0.0027	0.9962
0.31	9300	0.0025	0.0026	0.9962
0.3133	9400	0.0031	0.0029	0.9956
0.3167	9500	0.0041	0.0033	0.9949
0.32	9600	0.003	0.0028	0.9962
0.3233	9700	0.0029	0.0028	0.9960
0.3267	9800	0.0029	0.0029	0.9957
0.33	9900	0.0027	0.0028	0.9960
0.3333	10000	0.0036	0.0028	0.9963
0.3367	10100	0.0039	0.0027	0.9962
0.34	10200	0.0038	0.0029	0.9958
0.3433	10300	0.0047	0.0026	0.9962
0.3467	10400	0.0045	0.0027	0.9962
0.35	10500	0.0023	0.0026	0.9962
0.3533	10600	0.0035	0.0027	0.9963
0.3567	10700	0.0028	0.0028	0.9960
0.36	10800	0.0025	0.0027	0.9963
0.3633	10900	0.0035	0.0029	0.9957
0.3667	11000	0.0028	0.0028	0.9962
0.37	11100	0.0045	0.0027	0.9962
0.3733	11200	0.0032	0.0026	0.9965
0.3767	11300	0.0035	0.0026	0.9962
0.38	11400	0.005	0.0025	0.9965
0.3833	11500	0.0025	0.0025	0.9965
0.3867	11600	0.0034	0.0026	0.9963
0.39	11700	0.0035	0.0026	0.9963
0.3933	11800	0.0024	0.0029	0.9956
0.3967	11900	0.0034	0.0025	0.9965
0.4	12000	0.0036	0.0024	0.9968
0.4033	12100	0.003	0.0025	0.9968
0.4067	12200	0.0029	0.0025	0.9964
0.41	12300	0.0036	0.0025	0.9965
0.4133	12400	0.0016	0.0024	0.9966
0.4167	12500	0.0029	0.0025	0.9965
0.42	12600	0.0037	0.0024	0.9969
0.4233	12700	0.0025	0.0023	0.9968
0.4267	12800	0.0039	0.0022	0.9972
0.43	12900	0.0024	0.0022	0.9972
0.4333	13000	0.0038	0.0023	0.9968
0.4367	13100	0.0034	0.0022	0.9969
0.44	13200	0.0024	0.0023	0.9967
0.4433	13300	0.0026	0.0025	0.9964
0.4467	13400	0.0028	0.0024	0.9966
0.45	13500	0.0036	0.0024	0.9965
0.4533	13600	0.0025	0.0024	0.9965
0.4567	13700	0.0035	0.0024	0.9967
0.46	13800	0.0018	0.0023	0.9966
0.4633	13900	0.0028	0.0023	0.9968
0.4667	14000	0.0033	0.0022	0.9970
0.47	14100	0.0018	0.0022	0.9970
0.4733	14200	0.003	0.0021	0.9971
0.4767	14300	0.0021	0.0021	0.9971
0.48	14400	0.0029	0.0021	0.9971
0.4833	14500	0.0027	0.0023	0.9969
0.4867	14600	0.0023	0.0021	0.9971
0.49	14700	0.0026	0.0021	0.9972
0.4933	14800	0.0019	0.0021	0.9969
0.4967	14900	0.0024	0.0022	0.9968
0.5	15000	0.0025	0.0021	0.9968
0.5033	15100	0.0026	0.0021	0.9970
0.5067	15200	0.0019	0.0022	0.997
0.51	15300	0.0029	0.0023	0.9968
0.5133	15400	0.0026	0.0021	0.9970
0.5167	15500	0.0027	0.0021	0.9969
0.52	15600	0.0022	0.0022	0.9971
0.5233	15700	0.0026	0.0020	0.9973
0.5267	15800	0.0026	0.0021	0.9973
0.53	15900	0.0022	0.0020	0.9974
0.5333	16000	0.0039	0.0020	0.9975
0.5367	16100	0.0017	0.0020	0.9975
0.54	16200	0.0022	0.0020	0.9975
0.5433	16300	0.002	0.0019	0.9974
0.5467	16400	0.0033	0.0019	0.9975
0.55	16500	0.0032	0.0019	0.9974
0.5533	16600	0.0019	0.0020	0.9975
0.5567	16700	0.0027	0.0019	0.9974
0.56	16800	0.0027	0.0019	0.9973
0.5633	16900	0.0023	0.0018	0.9976
0.5667	17000	0.002	0.0018	0.9976
0.57	17100	0.0024	0.0019	0.9975
0.5733	17200	0.0021	0.0020	0.9973
0.5767	17300	0.0038	0.0019	0.9973
0.58	17400	0.0018	0.0018	0.9975
0.5833	17500	0.0031	0.0018	0.9977
0.5867	17600	0.0021	0.0018	0.9976
0.59	17700	0.0023	0.0019	0.9974
0.5933	17800	0.0031	0.0018	0.9975
0.5967	17900	0.002	0.0019	0.9975
0.6	18000	0.002	0.0018	0.9975
0.6033	18100	0.003	0.0019	0.9976
0.6067	18200	0.0023	0.0018	0.9977
0.61	18300	0.0029	0.0019	0.9975
0.6133	18400	0.0023	0.0018	0.9977
0.6167	18500	0.0017	0.0018	0.9977
0.62	18600	0.0022	0.0018	0.9977
0.6233	18700	0.0023	0.0018	0.9977
0.6267	18800	0.0021	0.0017	0.9978
0.63	18900	0.002	0.0017	0.9978
0.6333	19000	0.0028	0.0018	0.9978
0.6367	19100	0.0024	0.0017	0.9978
0.64	19200	0.0029	0.0017	0.9977
0.6433	19300	0.003	0.0017	0.9979
0.6467	19400	0.0027	0.0017	0.9978
0.65	19500	0.0032	0.0017	0.9978
0.6533	19600	0.0025	0.0017	0.9978
0.6567	19700	0.002	0.0017	0.9978
0.66	19800	0.0018	0.0017	0.9978
0.6633	19900	0.002	0.0018	0.9975
0.6667	20000	0.0029	0.0017	0.9978
0.67	20100	0.0012	0.0017	0.9978
0.6733	20200	0.0024	0.0016	0.9980
0.6767	20300	0.0027	0.0016	0.9980
0.68	20400	0.0023	0.0016	0.9980
0.6833	20500	0.0025	0.0016	0.9981
0.6867	20600	0.0018	0.0016	0.9981
0.69	20700	0.0015	0.0016	0.9981
0.6933	20800	0.0017	0.0015	0.9980
0.6967	20900	0.0027	0.0015	0.9981
0.7	21000	0.0016	0.0015	0.9982
0.7033	21100	0.002	0.0015	0.9983
0.7067	21200	0.002	0.0015	0.9983
0.71	21300	0.0022	0.0015	0.9984
0.7133	21400	0.0019	0.0015	0.9983
0.7167	21500	0.0025	0.0015	0.9981
0.72	21600	0.0025	0.0015	0.9981
0.7233	21700	0.0021	0.0015	0.9982
0.7267	21800	0.002	0.0015	0.9981
0.73	21900	0.0025	0.0015	0.9982
0.7333	22000	0.0021	0.0015	0.9982
0.7367	22100	0.0017	0.0015	0.9983
0.74	22200	0.0021	0.0015	0.9982
0.7433	22300	0.0026	0.0015	0.9981
0.7467	22400	0.0016	0.0015	0.9981
0.75	22500	0.0021	0.0014	0.9981
0.7533	22600	0.002	0.0015	0.9981
0.7567	22700	0.002	0.0014	0.9981
0.76	22800	0.0025	0.0014	0.9982
0.7633	22900	0.0022	0.0015	0.998
0.7667	23000	0.0022	0.0014	0.9981
0.77	23100	0.0017	0.0014	0.9982
0.7733	23200	0.0024	0.0014	0.9983
0.7767	23300	0.0021	0.0014	0.9981
0.78	23400	0.0018	0.0014	0.9982
0.7833	23500	0.0025	0.0014	0.9981
0.7867	23600	0.0025	0.0014	0.9981
0.79	23700	0.0015	0.0014	0.9981
0.7933	23800	0.0023	0.0014	0.9982
0.7967	23900	0.0028	0.0014	0.9981
0.8	24000	0.0022	0.0014	0.9981
0.8033	24100	0.0019	0.0014	0.9983
0.8067	24200	0.0021	0.0014	0.9982
0.81	24300	0.002	0.0013	0.9982
0.8133	24400	0.0015	0.0013	0.9982
0.8167	24500	0.0021	0.0013	0.9984
0.82	24600	0.0016	0.0013	0.9983
0.8233	24700	0.0016	0.0013	0.9983
0.8267	24800	0.0016	0.0014	0.9982
0.83	24900	0.0016	0.0013	0.9983
0.8333	25000	0.0012	0.0013	0.9982
0.8367	25100	0.0019	0.0013	0.9983
0.84	25200	0.0014	0.0013	0.9983
0.8433	25300	0.0024	0.0013	0.9983
0.8467	25400	0.0014	0.0013	0.9983
0.85	25500	0.0013	0.0013	0.9983
0.8533	25600	0.0017	0.0014	0.9983
0.8567	25700	0.0019	0.0014	0.9981
0.86	25800	0.003	0.0013	0.9983
0.8633	25900	0.0012	0.0013	0.9983
0.8667	26000	0.0023	0.0013	0.9983
0.87	26100	0.0017	0.0013	0.9983
0.8733	26200	0.0017	0.0013	0.9982
0.8767	26300	0.002	0.0013	0.9983
0.88	26400	0.0017	0.0013	0.9983
0.8833	26500	0.0017	0.0013	0.9982
0.8867	26600	0.0017	0.0013	0.9983
0.89	26700	0.0005	0.0013	0.9984
0.8933	26800	0.0014	0.0013	0.9983
0.8967	26900	0.0018	0.0013	0.9983
0.9	27000	0.0011	0.0013	0.9983
0.9033	27100	0.0012	0.0013	0.9983
0.9067	27200	0.0012	0.0013	0.9983
0.91	27300	0.0015	0.0013	0.9982
0.9133	27400	0.0015	0.0013	0.9983
0.9167	27500	0.0016	0.0013	0.9983
0.92	27600	0.0015	0.0012	0.9984
0.9233	27700	0.0015	0.0013	0.9984
0.9267	27800	0.0013	0.0012	0.9984
0.93	27900	0.0021	0.0012	0.9984
0.9333	28000	0.0008	0.0013	0.9984
0.9367	28100	0.002	0.0013	0.9983
0.94	28200	0.0024	0.0012	0.9984
0.9433	28300	0.0018	0.0012	0.9984
0.9467	28400	0.001	0.0012	0.9984
0.95	28500	0.001	0.0012	0.9983
0.9533	28600	0.002	0.0012	0.9984
0.9567	28700	0.0019	0.0012	0.9984
0.96	28800	0.0012	0.0012	0.9984
0.9633	28900	0.0017	0.0012	0.9984
0.9667	29000	0.0018	0.0012	0.9984
0.97	29100	0.0015	0.0012	0.9984
0.9733	29200	0.0012	0.0012	0.9983
0.9767	29300	0.0021	0.0012	0.9984
0.98	29400	0.0015	0.0012	0.9983
0.9833	29500	0.0013	0.0012	0.9983
0.9867	29600	0.0012	0.0012	0.9983
0.99	29700	0.0017	0.0012	0.9983
0.9933	29800	0.0016	0.0012	0.9983
0.9967	29900	0.0011	0.0012	0.9983
1.0	30000	0.0017	0.0012	0.9984

Framework Versions

• Python: 3.11.1
• Sentence Transformers: 3.3.1
• Transformers: 4.47.0
• PyTorch: 2.1.1+cu121
• Accelerate: 1.2.0
• Datasets: 2.18.0
• 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",
}

TripletLoss

@misc{hermans2017defense,
    title={In Defense of the Triplet Loss for Person Re-Identification},
    author={Alexander Hermans and Lucas Beyer and Bastian Leibe},
    year={2017},
    eprint={1703.07737},
    archivePrefix={arXiv},
    primaryClass={cs.CV}
}