README.md

metadata

language:
  - en
license: apache-2.0
tags:
  - sentence-transformers
  - sparse-encoder
  - sparse
  - splade
  - generated_from_trainer
  - dataset_size:99000
  - loss:SpladeLoss
  - loss:SparseDistillKLDivLoss
  - loss:FlopsLoss
base_model: Luyu/co-condenser-marco
widget:
  - text: >-
      The ejection fraction may decrease if: 1  You have weakness of your heart
      muscle, such as dilated cardiomyopathy, which can be caused by a heart
      muscle problem, familial (genetic) cardiomyopathy, or systemic illnesses.
      2  A heart attack has damaged your heart.  You have problems with your
      heart's valves.
  - text: "One thing we avoided: Lots of alternative slime recipes swap Borax for liquid starch, shampoo, body wash, hand soap, contact lens solution, or laundry detergent. Those may seem benign â\x80\x94 and they might be â\x80\x94 but many of them contain derivatives or relatives of sodium borate too."
  - text: how do i get my mvr in pa
  - text: >-
      English is a language whose vocabulary is the composite of a surprising
      range of influences. We have pillaged words from Latin, Greek, Dutch,
      Arabic, Old Norse, Spanish, Italian, Hindi, and more besides to make
      English what it is today.
  - text: >-
      Weed Eater was a string trimmer company founded in 1971 in Houston, Texas
      by George C. Ballas, Sr. , the inventor of the device. The idea for the
      Weed Eater trimmer came to him from the spinning nylon bristles of an
      automatic car wash.He thought that he could come up with a similar
      technique to protect the bark on trees that he was trimming around. His
      company was eventually bought by Emerson Electric and merged with
      Poulan.Poulan/Weed Eater was later purchased by Electrolux, which spun off
      the outdoors division as Husqvarna AB in 2006.Inventor Ballas was the
      father of champion ballroom dancer Corky Ballas and the grandfather of
      Dancing with the Stars dancer Mark Ballas.George Ballas died on June 25,
      2011.he idea for the Weed Eater trimmer came to him from the spinning
      nylon bristles of an automatic car wash. He thought that he could come up
      with a similar technique to protect the bark on trees that he was trimming
      around. His company was eventually bought by Emerson Electric and merged
      with Poulan.
pipeline_tag: feature-extraction
library_name: sentence-transformers
metrics:
  - dot_accuracy@1
  - dot_accuracy@3
  - dot_accuracy@5
  - dot_accuracy@10
  - dot_precision@1
  - dot_precision@3
  - dot_precision@5
  - dot_precision@10
  - dot_recall@1
  - dot_recall@3
  - dot_recall@5
  - dot_recall@10
  - dot_ndcg@10
  - dot_mrr@10
  - dot_map@100
  - query_active_dims
  - query_sparsity_ratio
  - corpus_active_dims
  - corpus_sparsity_ratio
co2_eq_emissions:
  emissions: 80.19586757744133
  energy_consumed: 0.20631709530036207
  source: codecarbon
  training_type: fine-tuning
  on_cloud: false
  cpu_model: 13th Gen Intel(R) Core(TM) i7-13700K
  ram_total_size: 31.777088165283203
  hours_used: 0.606
  hardware_used: 1 x NVIDIA GeForce RTX 3090
model-index:
  - name: CoCondenser finetuned on MS MARCO
    results:
      - task:
          type: sparse-information-retrieval
          name: Sparse Information Retrieval
        dataset:
          name: NanoMSMARCO
          type: NanoMSMARCO
        metrics:
          - type: dot_accuracy@1
            value: 0.34
            name: Dot Accuracy@1
          - type: dot_accuracy@3
            value: 0.54
            name: Dot Accuracy@3
          - type: dot_accuracy@5
            value: 0.68
            name: Dot Accuracy@5
          - type: dot_accuracy@10
            value: 0.8
            name: Dot Accuracy@10
          - type: dot_precision@1
            value: 0.34
            name: Dot Precision@1
          - type: dot_precision@3
            value: 0.18
            name: Dot Precision@3
          - type: dot_precision@5
            value: 0.136
            name: Dot Precision@5
          - type: dot_precision@10
            value: 0.08
            name: Dot Precision@10
          - type: dot_recall@1
            value: 0.34
            name: Dot Recall@1
          - type: dot_recall@3
            value: 0.54
            name: Dot Recall@3
          - type: dot_recall@5
            value: 0.68
            name: Dot Recall@5
          - type: dot_recall@10
            value: 0.8
            name: Dot Recall@10
          - type: dot_ndcg@10
            value: 0.5536387082079334
            name: Dot Ndcg@10
          - type: dot_mrr@10
            value: 0.47621428571428565
            name: Dot Mrr@10
          - type: dot_map@100
            value: 0.4871698934837092
            name: Dot Map@100
          - type: query_active_dims
            value: 224.6999969482422
            name: Query Active Dims
          - type: query_sparsity_ratio
            value: 0.9926380972102666
            name: Query Sparsity Ratio
          - type: corpus_active_dims
            value: 519.2117919921875
            name: Corpus Active Dims
          - type: corpus_sparsity_ratio
            value: 0.9829889328355879
            name: Corpus Sparsity Ratio
      - task:
          type: sparse-information-retrieval
          name: Sparse Information Retrieval
        dataset:
          name: NanoNFCorpus
          type: NanoNFCorpus
        metrics:
          - type: dot_accuracy@1
            value: 0.42
            name: Dot Accuracy@1
          - type: dot_accuracy@3
            value: 0.6
            name: Dot Accuracy@3
          - type: dot_accuracy@5
            value: 0.66
            name: Dot Accuracy@5
          - type: dot_accuracy@10
            value: 0.72
            name: Dot Accuracy@10
          - type: dot_precision@1
            value: 0.42
            name: Dot Precision@1
          - type: dot_precision@3
            value: 0.36666666666666664
            name: Dot Precision@3
          - type: dot_precision@5
            value: 0.3440000000000001
            name: Dot Precision@5
          - type: dot_precision@10
            value: 0.272
            name: Dot Precision@10
          - type: dot_recall@1
            value: 0.04202378573043321
            name: Dot Recall@1
          - type: dot_recall@3
            value: 0.09903115832887796
            name: Dot Recall@3
          - type: dot_recall@5
            value: 0.12166389615470453
            name: Dot Recall@5
          - type: dot_recall@10
            value: 0.14686804014722007
            name: Dot Recall@10
          - type: dot_ndcg@10
            value: 0.34309454902458447
            name: Dot Ndcg@10
          - type: dot_mrr@10
            value: 0.5205555555555555
            name: Dot Mrr@10
          - type: dot_map@100
            value: 0.16297327824481755
            name: Dot Map@100
          - type: query_active_dims
            value: 193.16000366210938
            name: Query Active Dims
          - type: query_sparsity_ratio
            value: 0.9936714499815835
            name: Query Sparsity Ratio
          - type: corpus_active_dims
            value: 712.2963256835938
            name: Corpus Active Dims
          - type: corpus_sparsity_ratio
            value: 0.9766628554588954
            name: Corpus Sparsity Ratio
      - task:
          type: sparse-information-retrieval
          name: Sparse Information Retrieval
        dataset:
          name: NanoNQ
          type: NanoNQ
        metrics:
          - type: dot_accuracy@1
            value: 0.44
            name: Dot Accuracy@1
          - type: dot_accuracy@3
            value: 0.66
            name: Dot Accuracy@3
          - type: dot_accuracy@5
            value: 0.78
            name: Dot Accuracy@5
          - type: dot_accuracy@10
            value: 0.84
            name: Dot Accuracy@10
          - type: dot_precision@1
            value: 0.44
            name: Dot Precision@1
          - type: dot_precision@3
            value: 0.22
            name: Dot Precision@3
          - type: dot_precision@5
            value: 0.16399999999999998
            name: Dot Precision@5
          - type: dot_precision@10
            value: 0.092
            name: Dot Precision@10
          - type: dot_recall@1
            value: 0.42
            name: Dot Recall@1
          - type: dot_recall@3
            value: 0.63
            name: Dot Recall@3
          - type: dot_recall@5
            value: 0.74
            name: Dot Recall@5
          - type: dot_recall@10
            value: 0.82
            name: Dot Recall@10
          - type: dot_ndcg@10
            value: 0.6291291692569377
            name: Dot Ndcg@10
          - type: dot_mrr@10
            value: 0.5775555555555555
            name: Dot Mrr@10
          - type: dot_map@100
            value: 0.5657484123460346
            name: Dot Map@100
          - type: query_active_dims
            value: 240.1199951171875
            name: Query Active Dims
          - type: query_sparsity_ratio
            value: 0.992132887913073
            name: Query Sparsity Ratio
          - type: corpus_active_dims
            value: 602.9705810546875
            name: Corpus Active Dims
          - type: corpus_sparsity_ratio
            value: 0.9802447224606944
            name: Corpus Sparsity Ratio
      - task:
          type: sparse-nano-beir
          name: Sparse Nano BEIR
        dataset:
          name: NanoBEIR mean
          type: NanoBEIR_mean
        metrics:
          - type: dot_accuracy@1
            value: 0.39999999999999997
            name: Dot Accuracy@1
          - type: dot_accuracy@3
            value: 0.6000000000000001
            name: Dot Accuracy@3
          - type: dot_accuracy@5
            value: 0.7066666666666667
            name: Dot Accuracy@5
          - type: dot_accuracy@10
            value: 0.7866666666666666
            name: Dot Accuracy@10
          - type: dot_precision@1
            value: 0.39999999999999997
            name: Dot Precision@1
          - type: dot_precision@3
            value: 0.25555555555555554
            name: Dot Precision@3
          - type: dot_precision@5
            value: 0.2146666666666667
            name: Dot Precision@5
          - type: dot_precision@10
            value: 0.14800000000000002
            name: Dot Precision@10
          - type: dot_recall@1
            value: 0.2673412619101444
            name: Dot Recall@1
          - type: dot_recall@3
            value: 0.423010386109626
            name: Dot Recall@3
          - type: dot_recall@5
            value: 0.5138879653849014
            name: Dot Recall@5
          - type: dot_recall@10
            value: 0.5889560133824067
            name: Dot Recall@10
          - type: dot_ndcg@10
            value: 0.5086208088298185
            name: Dot Ndcg@10
          - type: dot_mrr@10
            value: 0.5247751322751322
            name: Dot Mrr@10
          - type: dot_map@100
            value: 0.4052971946915204
            name: Dot Map@100
          - type: query_active_dims
            value: 219.32666524251303
            name: Query Active Dims
          - type: query_sparsity_ratio
            value: 0.9928141450349744
            name: Query Sparsity Ratio
          - type: corpus_active_dims
            value: 595.3305956849517
            name: Corpus Active Dims
          - type: corpus_sparsity_ratio
            value: 0.9804950332322604
            name: Corpus Sparsity Ratio

CoCondenser finetuned on MS MARCO

This is a SPLADE Sparse Encoder model finetuned from Luyu/co-condenser-marco using the sentence-transformers library. It maps sentences & paragraphs to a 30522-dimensional sparse vector space and can be used for semantic search and sparse retrieval.

Model Details

Model Description

• Model Type: SPLADE Sparse Encoder
• Base model: Luyu/co-condenser-marco
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 30522 dimensions
• Similarity Function: Dot Product
• Language: en
• License: apache-2.0

Model Sources

• Documentation: Sentence Transformers Documentation
• Documentation: Sparse Encoder Documentation
• Repository: Sentence Transformers on GitHub
• Hugging Face: Sparse Encoders on Hugging Face

Full Model Architecture

SparseEncoder(
  (0): MLMTransformer({'max_seq_length': 512, 'do_lower_case': False, 'architecture': 'BertForMaskedLM'})
  (1): SpladePooling({'pooling_strategy': 'max', 'activation_function': 'relu', 'word_embedding_dimension': 30522})
)

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 SparseEncoder

# Download from the 🤗 Hub
model = SparseEncoder("tomaarsen/splade-cocondenser-msmarco-kldiv-minilm-threshold")
# Run inference
queries = [
    "who started gladiator lacrosse",
]
documents = [
    'Weed Eater was a string trimmer company founded in 1971 in Houston, Texas by George C. Ballas, Sr. , the inventor of the device. The idea for the Weed Eater trimmer came to him from the spinning nylon bristles of an automatic car wash.He thought that he could come up with a similar technique to protect the bark on trees that he was trimming around. His company was eventually bought by Emerson Electric and merged with Poulan.Poulan/Weed Eater was later purchased by Electrolux, which spun off the outdoors division as Husqvarna AB in 2006.Inventor Ballas was the father of champion ballroom dancer Corky Ballas and the grandfather of Dancing with the Stars dancer Mark Ballas.George Ballas died on June 25, 2011.he idea for the Weed Eater trimmer came to him from the spinning nylon bristles of an automatic car wash. He thought that he could come up with a similar technique to protect the bark on trees that he was trimming around. His company was eventually bought by Emerson Electric and merged with Poulan.',
    "The earliest types of gladiator were named after Rome's enemies of that time: the Samnite, Thracian and Gaul. The Samnite, heavily armed, elegantly helmed and probably the most popular type, was renamed Secutor and the Gaul renamed Murmillo, once these former enemies had been conquered then absorbed into Rome's Empire.",
    'Summit Hill, PA. Sponsored Topics. Summit Hill is a borough in Carbon County, Pennsylvania, United States. The population was 2,974 at the 2000 census. Summit Hill is located at 40°49â\x80²39â\x80³N 75°51â\x80²57â\x80³W / 40.8275°N 75.86583°W / 40.8275; -75.86583 (40.827420, -75.865892).',
]
query_embeddings = model.encode_query(queries)
document_embeddings = model.encode_document(documents)
print(query_embeddings.shape, document_embeddings.shape)
# [1, 30522] [3, 30522]

# Get the similarity scores for the embeddings
similarities = model.similarity(query_embeddings, document_embeddings)
print(similarities)
# tensor([[ 83.3579, 166.2595,  13.1489]])

Evaluation

Metrics

Sparse Information Retrieval

• Datasets: NanoMSMARCO, NanoNFCorpus and NanoNQ
• Evaluated with SparseInformationRetrievalEvaluator

Metric	NanoMSMARCO	NanoNFCorpus	NanoNQ
dot_accuracy@1	0.34	0.42	0.44
dot_accuracy@3	0.54	0.6	0.66
dot_accuracy@5	0.68	0.66	0.78
dot_accuracy@10	0.8	0.72	0.84
dot_precision@1	0.34	0.42	0.44
dot_precision@3	0.18	0.3667	0.22
dot_precision@5	0.136	0.344	0.164
dot_precision@10	0.08	0.272	0.092
dot_recall@1	0.34	0.042	0.42
dot_recall@3	0.54	0.099	0.63
dot_recall@5	0.68	0.1217	0.74
dot_recall@10	0.8	0.1469	0.82
dot_ndcg@10	0.5536	0.3431	0.6291
dot_mrr@10	0.4762	0.5206	0.5776
dot_map@100	0.4872	0.163	0.5657
query_active_dims	224.7	193.16	240.12
query_sparsity_ratio	0.9926	0.9937	0.9921
corpus_active_dims	519.2118	712.2963	602.9706
corpus_sparsity_ratio	0.983	0.9767	0.9802

Sparse Nano BEIR

• Dataset: NanoBEIR_mean
• Evaluated with SparseNanoBEIREvaluator with these parameters:

  {
      "dataset_names": [
          "msmarco",
          "nfcorpus",
          "nq"
      ]
  }
  

Metric	Value
dot_accuracy@1	0.4
dot_accuracy@3	0.6
dot_accuracy@5	0.7067
dot_accuracy@10	0.7867
dot_precision@1	0.4
dot_precision@3	0.2556
dot_precision@5	0.2147
dot_precision@10	0.148
dot_recall@1	0.2673
dot_recall@3	0.423
dot_recall@5	0.5139
dot_recall@10	0.589
dot_ndcg@10	0.5086
dot_mrr@10	0.5248
dot_map@100	0.4053
query_active_dims	219.3267
query_sparsity_ratio	0.9928
corpus_active_dims	595.3306
corpus_sparsity_ratio	0.9805

Training Details

Training Dataset

Unnamed Dataset

• Size: 99,000 training samples
• Columns: query, positive, negative, and label
• Approximate statistics based on the first 1000 samples:

  	query	positive	negative	label
  type	string	string	string	list
  details	min: 4 tokens mean: 9.2 tokens max: 34 tokens	min: 18 tokens mean: 79.86 tokens max: 219 tokens	min: 18 tokens mean: 79.96 tokens max: 270 tokens	size: 2 elements

• Samples:

  query	positive	negative	label
  rtn tv network	Home Shopping Network. Home Shopping Network (HSN) is an American broadcast, basic cable and satellite television network that is owned by HSN, Inc. (NASDAQ: HSNI), which also owns catalog company Cornerstone Brands. Based in St. Petersburg, Florida, United States, the home shopping channel has former and current sister channels in several other countries.	The Public Switched Telephone Network - The public switched telephone network (PSTN) is the international network of circuit-switched telephones. Learn more about PSTN at HowStuffWorks. x	[-1.0804121494293213, -5.908488750457764]
  how did president nixon react to the watergate investigation?	The Watergate scandal was a major political scandal that occurred in the United States during the early 1970s, following a break-in by five men at the Democratic National Committee headquarters at the Watergate office complex in Washington, D.C. on June 17, 1972, and President Richard Nixon's administration's subsequent attempt to cover up its involvement. After the five burglars were caught and the conspiracy was discovered, Watergate was investigated by the United States Congress. Meanwhile, N	The release of the tape was ordered by the Supreme Court on July 24, 1974, in a case known as United States v. Nixon. The court’s decision was unanimous. President Nixon released the tape on August 5. It was one of three conversations he had with Haldeman six days after the Watergate break-in. The tapes prove that he ordered a cover-up of the Watergate burglary. The Smoking Gun tape reveals that Nixon ordered the FBI to abandon its investigation of the break-in. [Read more…]	[4.117279052734375, 3.191757917404175]
  what is a summary offense in pennsylvania	We provide cost effective house arrest and electronic monitoring services to magisterial district court systems throughout Pennsylvania including York, Harrisburg, Philadelphia and Allentown.In addition, we also serve the York County, Lancaster County and Chester County.e provide cost effective house arrest and electronic monitoring services to magisterial district court systems throughout Pennsylvania including York, Harrisburg, Philadelphia and Allentown.	In order to be convicted of Simple Assault, one must cause bodily injury. To be convicted of Aggravated Assault, one must cause serious bodily injury. From my research, Pennsylvania law defines bodily injury as the impairment of physical condition or substantial pain.	[-8.954689025878906, -1.3361705541610718]

• Loss: SpladeLoss with these parameters:

  {
      "loss": "SparseDistillKLDivLoss",
      "lambda_corpus": 0.0005,
      "lambda_query": 0.0005,
      "corpus_threshold": 200,
      "query_threshold": 50
  }
  

Evaluation Dataset

Unnamed Dataset

• Size: 1,000 evaluation samples
• Columns: query, positive, negative, and label
• Approximate statistics based on the first 1000 samples:

  	query	positive	negative	label
  type	string	string	string	list
  details	min: 4 tokens mean: 9.12 tokens max: 37 tokens	min: 17 tokens mean: 78.91 tokens max: 239 tokens	min: 25 tokens mean: 81.25 tokens max: 239 tokens	size: 2 elements

• Samples:

  query	positive	negative	label
  how long to cook roast beef for	Roasting times for beef. Preheat your oven to 160°C (325°F) and use these cooking times to prepare a roast that's moist, tender and delicious. Your roast should be covered with foil for the first half of the roasting time to prevent drying the outer layer.3 to 5lb Joint 1½ to 2 hours.reheat your oven to 160°C (325°F) and use these cooking times to prepare a roast that's moist, tender and delicious. Your roast should be covered with foil for the first half of the roasting time to prevent drying the outer layer.	Estimating Cooking Time for Large Beef Roasts. If you roast at a steady 325F (160C), subtract 2 minutes or so per pound. If the roast is refrigerated just before going into the oven, add 2 or 3 minutes per pound. WARNING NOTES: Remember, the rib roast will continue to cook as it sets.	[6.501978874206543, 8.214995384216309]
  definition of fire inspection	Learn how to do a monthly fire extinguisher inspection in your workplace. Departments must assign an individual to inspect monthly the extinguishers in or adjacent to the department's facilities.1 Read Fire Extinguisher Types and Maintenance for more information.earn how to do a monthly fire extinguisher inspection in your workplace. Departments must assign an individual to inspect monthly the extinguishers in or adjacent to the department's facilities.	reconnaissance by fire-a method of reconnaissance in which fire is placed on a suspected enemy position in order to cause the enemy to disclose his presence by moving or returning fire. reconnaissance in force-an offensive operation designed to discover or test the enemy's strength (or to obtain other information). mission undertaken to obtain, by visual observation or other detection methods, information about the activities and resources of an enemy or potential enemy, or to secure data concerning the meteorological, hydrographic, or geographic characteristics of a particular area.	[-0.38299351930618286, -0.9372650384902954]
  how many stores does family dollar have	Property Spotlight: New Retail Center at Hamilton & Warner - Outlots Available!! Family Dollar is closing stores following a disappointing second quarter. Family Dollar Stores Inc. won’t just be cutting prices in an attempt to boost its business – it’ll be closing stores as well. The Matthews, N.C.-based discount retailer plans to shutter 370 under-performing shops, according to the Charlotte Business Journal.	Glassdoor has 1,976 Family Dollar Stores reviews submitted anonymously by Family Dollar Stores employees. Read employee reviews and ratings on Glassdoor to decide if Family Dollar Stores is right for you.	[4.726407527923584, 8.284608840942383]

• Loss: SpladeLoss with these parameters:

  {
      "loss": "SparseDistillKLDivLoss",
      "lambda_corpus": 0.0005,
      "lambda_query": 0.0005,
      "corpus_threshold": 200,
      "query_threshold": 50
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 16
• per_device_eval_batch_size: 16
• learning_rate: 2e-05
• 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: 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: 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
• 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
• router_mapping: {}
• learning_rate_mapping: {}

Training Logs

Epoch	Step	Training Loss	Validation Loss	NanoMSMARCO_dot_ndcg@10	NanoNFCorpus_dot_ndcg@10	NanoNQ_dot_ndcg@10	NanoBEIR_mean_dot_ndcg@10
0.0162	100	21.4806	-	-	-	-	-
0.0323	200	2.4153	-	-	-	-	-
0.0485	300	0.9375	-	-	-	-	-
0.0646	400	0.8705	-	-	-	-	-
0.0808	500	0.8	0.7651	0.4601	0.2942	0.5525	0.4356
0.0970	600	0.8068	-	-	-	-	-
0.1131	700	0.7366	-	-	-	-	-
0.1293	800	0.7355	-	-	-	-	-
0.1454	900	0.731	-	-	-	-	-
0.1616	1000	0.7211	0.6796	0.5755	0.3193	0.5985	0.4978
0.1778	1100	0.7004	-	-	-	-	-
0.1939	1200	0.6903	-	-	-	-	-
0.2101	1300	0.67	-	-	-	-	-
0.2262	1400	0.6922	-	-	-	-	-
0.2424	1500	0.6474	0.6541	0.5409	0.3013	0.5737	0.4720
0.2586	1600	0.6538	-	-	-	-	-
0.2747	1700	0.6579	-	-	-	-	-
0.2909	1800	0.6577	-	-	-	-	-
0.3070	1900	0.6685	-	-	-	-	-
0.3232	2000	0.6298	0.6256	0.5603	0.3112	0.6109	0.4941
0.3394	2100	0.6591	-	-	-	-	-
0.3555	2200	0.6383	-	-	-	-	-
0.3717	2300	0.6601	-	-	-	-	-
0.3878	2400	0.6368	-	-	-	-	-
0.4040	2500	0.6531	0.5789	0.5880	0.3244	0.6298	0.5141
0.4202	2600	0.6351	-	-	-	-	-
0.4363	2700	0.5998	-	-	-	-	-
0.4525	2800	0.6148	-	-	-	-	-
0.4686	2900	0.61	-	-	-	-	-
0.4848	3000	0.6052	0.5768	0.5463	0.3064	0.6269	0.4932
0.5010	3100	0.5946	-	-	-	-	-
0.5171	3200	0.5907	-	-	-	-	-
0.5333	3300	0.6217	-	-	-	-	-
0.5495	3400	0.5938	-	-	-	-	-
0.5656	3500	0.5987	0.5290	0.5763	0.3221	0.6655	0.5213
0.5818	3600	0.5757	-	-	-	-	-
0.5979	3700	0.5786	-	-	-	-	-
0.6141	3800	0.5685	-	-	-	-	-
0.6303	3900	0.5623	-	-	-	-	-
0.6464	4000	0.5671	0.5241	0.5798	0.3329	0.6248	0.5125
0.6626	4100	0.5847	-	-	-	-	-
0.6787	4200	0.5855	-	-	-	-	-
0.6949	4300	0.5822	-	-	-	-	-
0.7111	4400	0.5497	-	-	-	-	-
0.7272	4500	0.5552	0.5060	0.5700	0.3531	0.6304	0.5178
0.7434	4600	0.5984	-	-	-	-	-
0.7595	4700	0.5437	-	-	-	-	-
0.7757	4800	0.5338	-	-	-	-	-
0.7919	4900	0.5728	-	-	-	-	-
0.8080	5000	0.5314	0.5116	0.5729	0.3386	0.6245	0.5120
0.8242	5100	0.544	-	-	-	-	-
0.8403	5200	0.5305	-	-	-	-	-
0.8565	5300	0.5667	-	-	-	-	-
0.8727	5400	0.5621	-	-	-	-	-
0.8888	5500	0.5407	0.4935	0.5589	0.3436	0.6325	0.5116
0.9050	5600	0.5586	-	-	-	-	-
0.9211	5700	0.522	-	-	-	-	-
0.9373	5800	0.5255	-	-	-	-	-
0.9535	5900	0.5338	-	-	-	-	-
0.9696	6000	0.5665	0.4871	0.5489	0.3415	0.6274	0.5059
0.9858	6100	0.558	-	-	-	-	-
-1	-1	-	-	0.5536	0.3431	0.6291	0.5086

Environmental Impact

Carbon emissions were measured using CodeCarbon.

• Energy Consumed: 0.206 kWh
• Carbon Emitted: 0.080 kg of CO2
• Hours Used: 0.606 hours

Training Hardware

• On Cloud: No
• GPU Model: 1 x NVIDIA GeForce RTX 3090
• CPU Model: 13th Gen Intel(R) Core(TM) i7-13700K
• RAM Size: 31.78 GB

Framework Versions

• Python: 3.11.6
• Sentence Transformers: 4.2.0.dev0
• Transformers: 4.52.4
• PyTorch: 2.7.1+cu126
• Accelerate: 1.5.1
• Datasets: 2.21.0
• Tokenizers: 0.21.1

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",
}

SpladeLoss

@misc{formal2022distillationhardnegativesampling,
      title={From Distillation to Hard Negative Sampling: Making Sparse Neural IR Models More Effective},
      author={Thibault Formal and Carlos Lassance and Benjamin Piwowarski and Stéphane Clinchant},
      year={2022},
      eprint={2205.04733},
      archivePrefix={arXiv},
      primaryClass={cs.IR},
      url={https://arxiv.org/abs/2205.04733},
}

SparseDistillKLDivLoss

@misc{lin2020distillingdenserepresentationsranking,
      title={Distilling Dense Representations for Ranking using Tightly-Coupled Teachers},
      author={Sheng-Chieh Lin and Jheng-Hong Yang and Jimmy Lin},
      year={2020},
      eprint={2010.11386},
      archivePrefix={arXiv},
      primaryClass={cs.IR},
      url={https://arxiv.org/abs/2010.11386},
}

FlopsLoss

@article{paria2020minimizing,
    title={Minimizing flops to learn efficient sparse representations},
    author={Paria, Biswajit and Yeh, Chih-Kuan and Yen, Ian EH and Xu, Ning and Ravikumar, Pradeep and P{'o}czos, Barnab{'a}s},
    journal={arXiv preprint arXiv:2004.05665},
    year={2020}
}