lv12/esci-nomic-embed-text-v1_5_1

SentenceTransformer based on nomic-ai/nomic-embed-text-v1.5

This is a sentence-transformers model finetuned from nomic-ai/nomic-embed-text-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: nomic-ai/nomic-embed-text-v1.5
• Maximum Sequence Length: 8192 tokens
• Output Dimensionality: 768 tokens
• Similarity Function: Cosine Similarity

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: NomicBertModel 
  (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})
)

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("sentence_transformers_model_id")
# Run inference
sentences = [
    'search_query: dab rig',
    'search_query: volcano weed vaporizer',
    'search_query: 22 gold chain for men',
]
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

Triplet

• Dataset: triplet-esci
• Evaluated with TripletEvaluator

Metric	Value
cosine_accuracy	0.7405
dot_accuracy	0.269
manhattan_accuracy	0.7432
euclidean_accuracy	0.7457
max_accuracy	0.7457

Training Details

Training Dataset

Unnamed Dataset

• Size: 167,039 training samples
• Columns: anchor, positive, and negative
• Approximate statistics based on the first 1000 samples:

  	anchor	positive	negative
  type	string	string	string
  details	min: 7 tokens mean: 11.1 tokens max: 38 tokens	min: 14 tokens mean: 43.23 tokens max: 124 tokens	min: 16 tokens mean: 43.16 tokens max: 97 tokens

• Samples:

  anchor	positive	negative
  search_query: foos ball coffee table	search_document: KICK Vanquish 55" in Foosball Table, KICK, Blue/Gray	search_document: KICK Legend 55" Foosball Table (Black), KICK, Black
  search_query: bathroom rugs white washable	search_document: Luxury Bath Mat Floor Towel Set - Absorbent Cotton Hotel Spa Shower/Bathtub Mats [Not a Bathroom Rug] 22"x34"	White
  search_query: kids gloves	search_document: EvridWear Boys Girls Magic Stretch Gripper Gloves 3 Pair Pack Assortment, Kids One Size Winter Warm Gloves Children (8-14Years, 3 Pairs Camo), Evridwear, 3 Pairs Camo	search_document: Body Glove Little Boys 2-Piece UPF 50+ Rash Guard Swimsuit Set (2 Piece), All Black, Size 5, Body Glove, All Black

• Loss: CachedMultipleNegativesRankingLoss with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim"
  }
  

Evaluation Dataset

Unnamed Dataset

• Size: 10,000 evaluation samples
• Columns: anchor, positive, and negative
• Approximate statistics based on the first 1000 samples:

  	anchor	positive	negative
  type	string	string	string
  details	min: 7 tokens mean: 11.44 tokens max: 31 tokens	min: 16 tokens mean: 42.26 tokens max: 92 tokens	min: 16 tokens mean: 42.28 tokens max: 105 tokens

• Samples:

  anchor	positive	negative
  search_query: defender series iphone 8	search_document: Hand-e Muscle Series Belt Clip Case for Apple iPhone 7 / iPhone 8 / iPhone SE “2020” (4.7”) 2-in-1 Protective Defender w Screen Protector & Holster & Kickstand/Shock & Drop Proof – Camouflage/Orange, Hand-e, Camouflage / Orange	search_document: OtterBox Defender Series Rugged Case for iPhone 8 PLUS & iPhone 7 PLUS - Case Only - Non-Retail Packaging - Dark Lake - With Microbial Defense, OtterBox, Dark Lake
  search_query: joy mangano	search_document: Joy by Joy Mangano 11-Piece Complete Luxury Towel Set, Ivory, Joy Mangano, Ivory	search_document: BAGSMART Jewelry Organizer Case Travel Jewelry Storage Bag for Necklace, Earrings, Rings, Bracelet, Soft Pink, BAGSMART, Soft Pink
  search_query: cashel fly masks for horses without ears	search_document: Cashel Crusader Designer Horse Fly Mask, Leopard, Weanling, Cashel, Leopard	search_document: Cashel Crusader Designer Horse Fly Mask with Ears, Teal Tribal, Weanling, Cashel, Teal Tribal

• Loss: CachedMultipleNegativesRankingLoss with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim"
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• per_device_train_batch_size: 4
• per_device_eval_batch_size: 4
• gradient_accumulation_steps: 4
• learning_rate: 1e-06
• num_train_epochs: 5
• lr_scheduler_type: cosine_with_restarts
• warmup_ratio: 0.1
• dataloader_drop_last: True
• dataloader_num_workers: 4
• dataloader_prefetch_factor: 2
• load_best_model_at_end: True
• batch_sampler: no_duplicates

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• prediction_loss_only: True
• per_device_train_batch_size: 4
• per_device_eval_batch_size: 4
• per_gpu_train_batch_size: None
• per_gpu_eval_batch_size: None
• gradient_accumulation_steps: 4
• eval_accumulation_steps: None
• learning_rate: 1e-06
• weight_decay: 0.0
• adam_beta1: 0.9
• adam_beta2: 0.999
• adam_epsilon: 1e-08
• max_grad_norm: 1.0
• num_train_epochs: 5
• max_steps: -1
• lr_scheduler_type: cosine_with_restarts
• 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
• 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: 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: True
• dataloader_num_workers: 4
• dataloader_prefetch_factor: 2
• 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}
• 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: False
• hub_always_push: False
• gradient_checkpointing: False
• gradient_checkpointing_kwargs: None
• include_inputs_for_metrics: False
• 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
• batch_sampler: no_duplicates
• multi_dataset_batch_sampler: proportional

Training Logs

Click to expand

Epoch	Step	Training Loss	loss	triplet-esci_cosine_accuracy
0.0096	100	0.6669	-	-
0.0192	200	0.6633	-	-
0.0287	300	0.6575	-	-
0.0383	400	0.6638	-	-
0.0479	500	0.6191	-	-
0.0575	600	0.6464	-	-
0.0671	700	0.6291	-	-
0.0766	800	0.5973	-	-
0.0862	900	0.605	-	-
0.0958	1000	0.6278	0.6525	0.7269
0.1054	1100	0.6041	-	-
0.1149	1200	0.6077	-	-
0.1245	1300	0.589	-	-
0.1341	1400	0.5811	-	-
0.1437	1500	0.5512	-	-
0.1533	1600	0.5907	-	-
0.1628	1700	0.5718	-	-
0.1724	1800	0.5446	-	-
0.1820	1900	0.546	-	-
0.1916	2000	0.5141	0.6105	0.7386
0.2012	2100	0.5359	-	-
0.2107	2200	0.5093	-	-
0.2203	2300	0.5384	-	-
0.2299	2400	0.5582	-	-
0.2395	2500	0.5038	-	-
0.2490	2600	0.5031	-	-
0.2586	2700	0.5393	-	-
0.2682	2800	0.4979	-	-
0.2778	2900	0.5221	-	-
0.2874	3000	0.4956	0.5852	0.7495
0.2969	3100	0.506	-	-
0.3065	3200	0.4962	-	-
0.3161	3300	0.4713	-	-
0.3257	3400	0.5016	-	-
0.3353	3500	0.4749	-	-
0.3448	3600	0.4732	-	-
0.3544	3700	0.4789	-	-
0.3640	3800	0.4825	-	-
0.3736	3900	0.4803	-	-
0.3832	4000	0.4471	0.5743	0.7546
0.3927	4100	0.4593	-	-
0.4023	4200	0.4481	-	-
0.4119	4300	0.4603	-	-
0.4215	4400	0.4569	-	-
0.4310	4500	0.4807	-	-
0.4406	4600	0.4368	-	-
0.4502	4700	0.4532	-	-
0.4598	4800	0.4432	-	-
0.4694	4900	0.4802	-	-
0.4789	5000	0.4643	0.5663	0.7593
0.4885	5100	0.4154	-	-
0.4981	5200	0.4441	-	-
0.5077	5300	0.4156	-	-
0.5173	5400	0.4273	-	-
0.5268	5500	0.3988	-	-
0.5364	5600	0.3942	-	-
0.5460	5700	0.4186	-	-
0.5556	5800	0.423	-	-
0.5651	5900	0.434	-	-
0.5747	6000	0.4136	0.5704	0.7616
0.5843	6100	0.3968	-	-
0.5939	6200	0.4045	-	-
0.6035	6300	0.4122	-	-
0.6130	6400	0.3618	-	-
0.6226	6500	0.341	-	-
0.6322	6600	0.3689	-	-
0.6418	6700	0.3621	-	-
0.6514	6800	0.3774	-	-
0.6609	6900	0.3519	-	-
0.6705	7000	0.3974	0.5729	0.7644
0.6801	7100	0.3443	-	-
0.6897	7200	0.3665	-	-
0.6993	7300	0.3683	-	-
0.7088	7400	0.3593	-	-
0.7184	7500	0.3419	-	-
0.7280	7600	0.3587	-	-
0.7376	7700	0.3463	-	-
0.7471	7800	0.3417	-	-
0.7567	7900	0.32	-	-
0.7663	8000	0.32	0.5735	0.7677
0.7759	8100	0.3296	-	-
0.7855	8200	0.3492	-	-
0.7950	8300	0.3022	-	-
0.8046	8400	0.3159	-	-
0.8142	8500	0.3172	-	-
0.8238	8600	0.3157	-	-
0.8334	8700	0.3271	-	-
0.8429	8800	0.337	-	-
0.8525	8900	0.322	-	-
0.8621	9000	0.3187	0.5803	0.7652
0.8717	9100	0.307	-	-
0.8812	9200	0.2984	-	-
0.8908	9300	0.2727	-	-
0.9004	9400	0.304	-	-
0.9100	9500	0.321	-	-
0.9196	9600	0.304	-	-
0.9291	9700	0.3302	-	-
0.9387	9800	0.3302	-	-
0.9483	9900	0.3134	-	-
0.9579	10000	0.2936	0.5858	0.7671
0.9675	10100	0.2953	-	-
0.9770	10200	0.3035	-	-
0.9866	10300	0.303	-	-
0.9962	10400	0.2606	-	-
1.0058	10500	0.2615	-	-
1.0153	10600	0.2703	-	-
1.0249	10700	0.2761	-	-
1.0345	10800	0.2559	-	-
1.0441	10900	0.2672	-	-
1.0537	11000	0.2656	0.5933	0.7676
1.0632	11100	0.2825	-	-
1.0728	11200	0.2484	-	-
1.0824	11300	0.2472	-	-
1.0920	11400	0.2678	-	-
1.1016	11500	0.2443	-	-
1.1111	11600	0.2685	-	-
1.1207	11700	0.2504	-	-
1.1303	11800	0.2431	-	-
1.1399	11900	0.2248	-	-
1.1495	12000	0.2229	0.5958	0.7688
1.1590	12100	0.228	-	-
1.1686	12200	0.2304	-	-
1.1782	12300	0.2193	-	-
1.1878	12400	0.2238	-	-
1.1973	12500	0.1957	-	-
1.2069	12600	0.2075	-	-
1.2165	12700	0.2014	-	-
1.2261	12800	0.2222	-	-
1.2357	12900	0.2059	-	-
1.2452	13000	0.2051	0.6077	0.7651
1.2548	13100	0.2076	-	-
1.2644	13200	0.226	-	-
1.2740	13300	0.1941	-	-
1.2836	13400	0.2053	-	-
1.2931	13500	0.2003	-	-
1.3027	13600	0.1947	-	-
1.3123	13700	0.1914	-	-
1.3219	13800	0.1956	-	-
1.3314	13900	0.1862	-	-
1.3410	14000	0.1873	0.6110	0.7646
1.3506	14100	0.1812	-	-
1.3602	14200	0.1828	-	-
1.3698	14300	0.1696	-	-
1.3793	14400	0.1705	-	-
1.3889	14500	0.1746	-	-
1.3985	14600	0.1756	-	-
1.4081	14700	0.1682	-	-
1.4177	14800	0.1769	-	-
1.4272	14900	0.1795	-	-
1.4368	15000	0.1736	0.6278	0.7616
1.4464	15100	0.1546	-	-
1.4560	15200	0.1643	-	-
1.4656	15300	0.1903	-	-
1.4751	15400	0.1902	-	-
1.4847	15500	0.1531	-	-
1.4943	15600	0.1711	-	-
1.5039	15700	0.1546	-	-
1.5134	15800	0.1503	-	-
1.5230	15900	0.1429	-	-
1.5326	16000	0.147	0.6306	0.7623
1.5422	16100	0.1507	-	-
1.5518	16200	0.152	-	-
1.5613	16300	0.1602	-	-
1.5709	16400	0.1541	-	-
1.5805	16500	0.1491	-	-
1.5901	16600	0.1378	-	-
1.5997	16700	0.1505	-	-
1.6092	16800	0.1334	-	-
1.6188	16900	0.1288	-	-
1.6284	17000	0.1168	0.6372	0.7629
1.6380	17100	0.135	-	-
1.6475	17200	0.1239	-	-
1.6571	17300	0.1398	-	-
1.6667	17400	0.1292	-	-
1.6763	17500	0.1414	-	-
1.6859	17600	0.116	-	-
1.6954	17700	0.1302	-	-
1.7050	17800	0.1194	-	-
1.7146	17900	0.1394	-	-
1.7242	18000	0.1316	0.6561	0.7592
1.7338	18100	0.1246	-	-
1.7433	18200	0.1277	-	-
1.7529	18300	0.1055	-	-
1.7625	18400	0.1211	-	-
1.7721	18500	0.1107	-	-
1.7817	18600	0.1145	-	-
1.7912	18700	0.1162	-	-
1.8008	18800	0.1114	-	-
1.8104	18900	0.1182	-	-
1.8200	19000	0.1152	0.6567	0.7591
1.8295	19100	0.1212	-	-
1.8391	19200	0.1253	-	-
1.8487	19300	0.115	-	-
1.8583	19400	0.1292	-	-
1.8679	19500	0.1151	-	-
1.8774	19600	0.1005	-	-
1.8870	19700	0.1079	-	-
1.8966	19800	0.0954	-	-
1.9062	19900	0.1045	-	-
1.9158	20000	0.1086	0.6727	0.7554
1.9253	20100	0.1174	-	-
1.9349	20200	0.1108	-	-
1.9445	20300	0.0992	-	-
1.9541	20400	0.1168	-	-
1.9636	20500	0.1028	-	-
1.9732	20600	0.1126	-	-
1.9828	20700	0.1113	-	-
1.9924	20800	0.1065	-	-
2.0020	20900	0.078	-	-
2.0115	21000	0.0921	0.6727	0.7568
2.0211	21100	0.0866	-	-
2.0307	21200	0.0918	-	-
2.0403	21300	0.0893	-	-
2.0499	21400	0.0882	-	-
2.0594	21500	0.0986	-	-
2.0690	21600	0.0923	-	-
2.0786	21700	0.0805	-	-
2.0882	21800	0.0887	-	-
2.0978	21900	0.1	-	-
2.1073	22000	0.0957	0.6854	0.7539
2.1169	22100	0.0921	-	-
2.1265	22200	0.0892	-	-
2.1361	22300	0.0805	-	-
2.1456	22400	0.0767	-	-
2.1552	22500	0.0715	-	-
2.1648	22600	0.083	-	-
2.1744	22700	0.0755	-	-
2.1840	22800	0.075	-	-
2.1935	22900	0.0724	-	-
2.2031	23000	0.0822	0.6913	0.7534
2.2127	23100	0.0623	-	-
2.2223	23200	0.0765	-	-
2.2319	23300	0.0755	-	-
2.2414	23400	0.0786	-	-
2.2510	23500	0.0651	-	-
2.2606	23600	0.081	-	-
2.2702	23700	0.0664	-	-
2.2797	23800	0.0906	-	-
2.2893	23900	0.0714	-	-
2.2989	24000	0.0703	0.6971	0.7536
2.3085	24100	0.0672	-	-
2.3181	24200	0.0754	-	-
2.3276	24300	0.0687	-	-
2.3372	24400	0.0668	-	-
2.3468	24500	0.0616	-	-
2.3564	24600	0.0693	-	-
2.3660	24700	0.0587	-	-
2.3755	24800	0.0612	-	-
2.3851	24900	0.0559	-	-
2.3947	25000	0.0676	0.7128	0.7497
2.4043	25100	0.0607	-	-
2.4139	25200	0.0727	-	-
2.4234	25300	0.0573	-	-
2.4330	25400	0.0717	-	-
2.4426	25500	0.0493	-	-
2.4522	25600	0.0558	-	-
2.4617	25700	0.0676	-	-
2.4713	25800	0.0757	-	-
2.4809	25900	0.0735	-	-
2.4905	26000	0.056	0.7044	0.7513
2.5001	26100	0.0687	-	-
2.5096	26200	0.0592	-	-
2.5192	26300	0.057	-	-
2.5288	26400	0.0444	-	-
2.5384	26500	0.0547	-	-
2.5480	26600	0.0605	-	-
2.5575	26700	0.066	-	-
2.5671	26800	0.0631	-	-
2.5767	26900	0.0634	-	-
2.5863	27000	0.0537	0.7127	0.7512
2.5958	27100	0.0535	-	-
2.6054	27200	0.0572	-	-
2.6150	27300	0.0473	-	-
2.6246	27400	0.0418	-	-
2.6342	27500	0.0585	-	-
2.6437	27600	0.0475	-	-
2.6533	27700	0.0549	-	-
2.6629	27800	0.0452	-	-
2.6725	27900	0.0514	-	-
2.6821	28000	0.0449	0.7337	0.7482
2.6916	28100	0.0544	-	-
2.7012	28200	0.041	-	-
2.7108	28300	0.0599	-	-
2.7204	28400	0.057	-	-
2.7300	28500	0.0503	-	-
2.7395	28600	0.0487	-	-
2.7491	28700	0.0503	-	-
2.7587	28800	0.0446	-	-
2.7683	28900	0.042	-	-
2.7778	29000	0.0501	0.7422	0.7469
2.7874	29100	0.0494	-	-
2.7970	29200	0.0423	-	-
2.8066	29300	0.0508	-	-
2.8162	29400	0.0459	-	-
2.8257	29500	0.0514	-	-
2.8353	29600	0.0484	-	-
2.8449	29700	0.0571	-	-
2.8545	29800	0.0558	-	-
2.8641	29900	0.0466	-	-
2.8736	30000	0.0465	0.7478	0.7447
2.8832	30100	0.0463	-	-
2.8928	30200	0.0362	-	-
2.9024	30300	0.0435	-	-
2.9119	30400	0.0419	-	-
2.9215	30500	0.046	-	-
2.9311	30600	0.0451	-	-
2.9407	30700	0.0458	-	-
2.9503	30800	0.052	-	-
2.9598	30900	0.0454	-	-
2.9694	31000	0.0433	0.7580	0.745
2.9790	31100	0.0438	-	-
2.9886	31200	0.0537	-	-
2.9982	31300	0.033	-	-
3.0077	31400	0.0384	-	-
3.0173	31500	0.0349	-	-
3.0269	31600	0.0365	-	-
3.0365	31700	0.0397	-	-
3.0460	31800	0.0396	-	-
3.0556	31900	0.0358	-	-
3.0652	32000	0.0443	0.7592	0.7454
3.0748	32100	0.0323	-	-
3.0844	32200	0.0418	-	-
3.0939	32300	0.0463	-	-
3.1035	32400	0.0397	-	-
3.1131	32500	0.0425	-	-
3.1227	32600	0.0406	-	-
3.1323	32700	0.0454	-	-
3.1418	32800	0.0287	-	-
3.1514	32900	0.0267	-	-
3.1610	33000	0.0341	0.7672	0.7431
3.1706	33100	0.0357	-	-
3.1802	33200	0.0322	-	-
3.1897	33300	0.0367	-	-
3.1993	33400	0.0419	-	-
3.2089	33500	0.0349	-	-
3.2185	33600	0.0327	-	-
3.2280	33700	0.0377	-	-
3.2376	33800	0.0353	-	-
3.2472	33900	0.0305	-	-
3.2568	34000	0.0362	0.7668	0.7463
3.2664	34100	0.0311	-	-
3.2759	34200	0.0405	-	-
3.2855	34300	0.0401	-	-
3.2951	34400	0.0361	-	-
3.3047	34500	0.0302	-	-
3.3143	34600	0.0379	-	-
3.3238	34700	0.03	-	-
3.3334	34800	0.039	-	-
3.3430	34900	0.0288	-	-
3.3526	35000	0.0318	0.7782	0.7436
3.3621	35100	0.0283	-	-
3.3717	35200	0.029	-	-
3.3813	35300	0.0287	-	-
3.3909	35400	0.0343	-	-
3.4005	35500	0.0326	-	-
3.4100	35600	0.031	-	-
3.4196	35700	0.0304	-	-
3.4292	35800	0.0314	-	-
3.4388	35900	0.0286	-	-
3.4484	36000	0.0229	0.7978	0.7428
3.4579	36100	0.0258	-	-
3.4675	36200	0.043	-	-
3.4771	36300	0.042	-	-
3.4867	36400	0.029	-	-
3.4963	36500	0.0343	-	-
3.5058	36600	0.0317	-	-
3.5154	36700	0.0307	-	-
3.5250	36800	0.0251	-	-
3.5346	36900	0.025	-	-
3.5441	37000	0.0309	0.8002	0.7446
3.5537	37100	0.031	-	-
3.5633	37200	0.0345	-	-
3.5729	37300	0.0332	-	-
3.5825	37400	0.0346	-	-
3.5920	37500	0.026	-	-
3.6016	37600	0.0293	-	-
3.6112	37700	0.0268	-	-
3.6208	37800	0.0264	-	-
3.6304	37900	0.0259	-	-
3.6399	38000	0.032	0.7896	0.7438
3.6495	38100	0.0246	-	-
3.6591	38200	0.0279	-	-
3.6687	38300	0.0274	-	-
3.6782	38400	0.0241	-	-
3.6878	38500	0.027	-	-
3.6974	38600	0.022	-	-
3.7070	38700	0.0305	-	-
3.7166	38800	0.0368	-	-
3.7261	38900	0.0304	-	-
3.7357	39000	0.0249	0.7978	0.7437
3.7453	39100	0.0312	-	-
3.7549	39200	0.0257	-	-
3.7645	39300	0.0273	-	-
3.7740	39400	0.0209	-	-
3.7836	39500	0.0298	-	-
3.7932	39600	0.0282	-	-
3.8028	39700	0.028	-	-
3.8124	39800	0.0279	-	-
3.8219	39900	0.0283	-	-
3.8315	40000	0.0239	0.7982	0.7424
3.8411	40100	0.0378	-	-
3.8507	40200	0.028	-	-
3.8602	40300	0.0321	-	-
3.8698	40400	0.0289	-	-
3.8794	40500	0.027	-	-
3.8890	40600	0.0224	-	-
3.8986	40700	0.0236	-	-
3.9081	40800	0.0267	-	-
3.9177	40900	0.0228	-	-
3.9273	41000	0.0322	0.8101	0.7415
3.9369	41100	0.0262	-	-
3.9465	41200	0.0276	-	-
3.9560	41300	0.0292	-	-
3.9656	41400	0.0278	-	-
3.9752	41500	0.0262	-	-
3.9848	41600	0.0306	-	-
3.9943	41700	0.0238	-	-
4.0039	41800	0.0165	-	-
4.0135	41900	0.0241	-	-
4.0231	42000	0.0211	0.8092	0.742
4.0327	42100	0.0257	-	-
4.0422	42200	0.0236	-	-
4.0518	42300	0.0254	-	-
4.0614	42400	0.0248	-	-
4.0710	42500	0.026	-	-
4.0806	42600	0.0245	-	-
4.0901	42700	0.0325	-	-
4.0997	42800	0.0209	-	-
4.1093	42900	0.033	-	-
4.1189	43000	0.0265	0.8105	0.7412
4.1285	43100	0.027	-	-
4.1380	43200	0.0208	-	-
4.1476	43300	0.0179	-	-
4.1572	43400	0.0194	-	-
4.1668	43500	0.0217	-	-
4.1763	43600	0.0212	-	-
4.1859	43700	0.0226	-	-
4.1955	43800	0.0252	-	-
4.2051	43900	0.0293	-	-
4.2147	44000	0.0216	0.8029	0.7414
4.2242	44100	0.029	-	-
4.2338	44200	0.0216	-	-
4.2434	44300	0.0251	-	-
4.2530	44400	0.018	-	-
4.2626	44500	0.025	-	-
4.2721	44600	0.0225	-	-
4.2817	44700	0.0303	-	-
4.2913	44800	0.028	-	-
4.3009	44900	0.0203	-	-
4.3104	45000	0.026	0.8081	0.7405

Framework Versions

• Python: 3.10.12
• Sentence Transformers: 3.0.0
• Transformers: 4.38.2
• PyTorch: 2.1.2+cu121
• Accelerate: 0.27.2
• Datasets: 2.19.1
• Tokenizers: 0.15.2

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",
}

CachedMultipleNegativesRankingLoss

@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}
}