README.md

metadata

language:
  - en
license: apache-2.0
tags:
  - sentence-transformers
  - sparse-encoder
  - sparse
  - asymmetric
  - inference-free
  - splade
  - generated_from_trainer
  - dataset_size:99000
  - loss:SpladeLoss
  - loss:SparseMultipleNegativesRankingLoss
  - loss:FlopsLoss
widget:
  - text: >-
      Rollin' (Limp Bizkit song) The music video was filmed atop the South Tower
      of the former World Trade Center in New York City. The introduction
      features Ben Stiller and Stephen Dorff mistaking Fred Durst for the valet
      and giving him the keys to their Bentley Azure. Also making a cameo is
      break dancer Mr. Wiggles. The rest of the video has several cuts to Durst
      and his bandmates hanging out of the Bentley as they drive about
      Manhattan. The song Ben Stiller is playing at the beginning is "My
      Generation" from the same album. The video also features scenes of Fred
      Durst with five girls dancing in a room. The video was filmed around the
      same time as the film Zoolander, which explains Stiller and Dorff's
      appearance. Fred Durst has a small cameo in that film.
  - text: >-
      Maze Runner: The Death Cure On April 22, 2017, the studio delayed the
      release date once again, to February 9, 2018, in order to allow more time
      for post-production; months later, on August 25, the studio moved the
      release forward two weeks.[17] The film will premiere on January 26, 2018
      in 3D, IMAX and IMAX 3D.[18][19]
  - text: who played the dj in the movie the warriors
  - text: >-
      Lionel Messi Born and raised in central Argentina, Messi was diagnosed
      with a growth hormone deficiency as a child. At age 13, he relocated to
      Spain to join Barcelona, who agreed to pay for his medical treatment.
      After a fast progression through Barcelona's youth academy, Messi made his
      competitive debut aged 17 in October 2004. Despite being injury-prone
      during his early career, he established himself as an integral player for
      the club within the next three years, finishing 2007 as a finalist for
      both the Ballon d'Or and FIFA World Player of the Year award, a feat he
      repeated the following year. His first uninterrupted campaign came in the
      2008–09 season, during which he helped Barcelona achieve the first
      treble in Spanish football. At 22 years old, Messi won the Ballon d'Or and
      FIFA World Player of the Year award by record voting margins.
  - text: >-
      Send In the Clowns "Send In the Clowns" is a song written by Stephen
      Sondheim for the 1973 musical A Little Night Music, an adaptation of
      Ingmar Bergman's film Smiles of a Summer Night. It is a ballad from Act
      Two, in which the character Desirée reflects on the ironies and
      disappointments of her life. Among other things, she looks back on an
      affair years earlier with the lawyer Fredrik, who was deeply in love with
      her but whose marriage proposals she had rejected. Meeting him after so
      long, she realizes she is in love with him and finally ready to marry him,
      but now it is he who rejects her: he is in an unconsummated marriage with
      a much younger woman. Desirée proposes marriage to rescue him from this
      situation, but he declines, citing his dedication to his bride. Reacting
      to his rejection, Desirée sings this song. The song is later reprised as a
      coda after Fredrik's young wife runs away with his son, and Fredrik is
      finally free to accept Desirée's offer.[1]
datasets:
  - sentence-transformers/natural-questions
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: 63.662764436204725
  energy_consumed: 0.16378296081883995
  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: 0.432
  hardware_used: 1 x NVIDIA GeForce RTX 3090
model-index:
  - name: >-
      Inference-free SPLADE bert-base-uncased trained on Natural-Questions
      tuples
    results:
      - task:
          type: sparse-information-retrieval
          name: Sparse Information Retrieval
        dataset:
          name: NanoMSMARCO
          type: NanoMSMARCO
        metrics:
          - type: dot_accuracy@1
            value: 0.24
            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: 0.78
            name: Dot Accuracy@10
          - type: dot_precision@1
            value: 0.24
            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: 0.078
            name: Dot Precision@10
          - type: dot_recall@1
            value: 0.24
            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: 0.78
            name: Dot Recall@10
          - type: dot_ndcg@10
            value: 0.5089651763164315
            name: Dot Ndcg@10
          - type: dot_mrr@10
            value: 0.4219047619047619
            name: Dot Mrr@10
          - type: dot_map@100
            value: 0.432147340250938
            name: Dot Map@100
          - type: query_active_dims
            value: 7.21999979019165
            name: Query Active Dims
          - type: query_sparsity_ratio
            value: 0.999763449322122
            name: Query Sparsity Ratio
          - type: corpus_active_dims
            value: 64.50605010986328
            name: Corpus Active Dims
          - type: corpus_sparsity_ratio
            value: 0.9978865719772668
            name: Corpus Sparsity Ratio
      - task:
          type: sparse-information-retrieval
          name: Sparse Information Retrieval
        dataset:
          name: NanoNFCorpus
          type: NanoNFCorpus
        metrics:
          - type: dot_accuracy@1
            value: 0.44
            name: Dot Accuracy@1
          - type: dot_accuracy@3
            value: 0.54
            name: Dot Accuracy@3
          - type: dot_accuracy@5
            value: 0.56
            name: Dot Accuracy@5
          - type: dot_accuracy@10
            value: 0.58
            name: Dot Accuracy@10
          - type: dot_precision@1
            value: 0.44
            name: Dot Precision@1
          - type: dot_precision@3
            value: 0.3733333333333333
            name: Dot Precision@3
          - type: dot_precision@5
            value: 0.324
            name: Dot Precision@5
          - type: dot_precision@10
            value: 0.24599999999999997
            name: Dot Precision@10
          - type: dot_recall@1
            value: 0.022479220566210704
            name: Dot Recall@1
          - type: dot_recall@3
            value: 0.07612273703265512
            name: Dot Recall@3
          - type: dot_recall@5
            value: 0.09256181151372667
            name: Dot Recall@5
          - type: dot_recall@10
            value: 0.11369159268450002
            name: Dot Recall@10
          - type: dot_ndcg@10
            value: 0.3093526962928894
            name: Dot Ndcg@10
          - type: dot_mrr@10
            value: 0.4906666666666667
            name: Dot Mrr@10
          - type: dot_map@100
            value: 0.12838279162115543
            name: Dot Map@100
          - type: query_active_dims
            value: 5.659999847412109
            name: Query Active Dims
          - type: query_sparsity_ratio
            value: 0.9998145599945151
            name: Query Sparsity Ratio
          - type: corpus_active_dims
            value: 81.60819244384766
            name: Corpus Active Dims
          - type: corpus_sparsity_ratio
            value: 0.9973262501656561
            name: Corpus Sparsity Ratio
      - task:
          type: sparse-information-retrieval
          name: Sparse Information Retrieval
        dataset:
          name: NanoNQ
          type: NanoNQ
        metrics:
          - type: dot_accuracy@1
            value: 0.38
            name: Dot Accuracy@1
          - type: dot_accuracy@3
            value: 0.54
            name: Dot Accuracy@3
          - type: dot_accuracy@5
            value: 0.62
            name: Dot Accuracy@5
          - type: dot_accuracy@10
            value: 0.66
            name: Dot Accuracy@10
          - type: dot_precision@1
            value: 0.38
            name: Dot Precision@1
          - type: dot_precision@3
            value: 0.18
            name: Dot Precision@3
          - type: dot_precision@5
            value: 0.128
            name: Dot Precision@5
          - type: dot_precision@10
            value: 0.068
            name: Dot Precision@10
          - type: dot_recall@1
            value: 0.38
            name: Dot Recall@1
          - type: dot_recall@3
            value: 0.52
            name: Dot Recall@3
          - type: dot_recall@5
            value: 0.61
            name: Dot Recall@5
          - type: dot_recall@10
            value: 0.64
            name: Dot Recall@10
          - type: dot_ndcg@10
            value: 0.5187972911837246
            name: Dot Ndcg@10
          - type: dot_mrr@10
            value: 0.4807460317460317
            name: Dot Mrr@10
          - type: dot_map@100
            value: 0.48967329128750314
            name: Dot Map@100
          - type: query_active_dims
            value: 10.319999694824219
            name: Query Active Dims
          - type: query_sparsity_ratio
            value: 0.9996618832417657
            name: Query Sparsity Ratio
          - type: corpus_active_dims
            value: 53.484806060791016
            name: Corpus Active Dims
          - type: corpus_sparsity_ratio
            value: 0.9982476637815088
            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: 0.35333333333333333
            name: Dot Accuracy@1
          - type: dot_accuracy@3
            value: 0.54
            name: Dot Accuracy@3
          - type: dot_accuracy@5
            value: 0.6200000000000001
            name: Dot Accuracy@5
          - type: dot_accuracy@10
            value: 0.6733333333333333
            name: Dot Accuracy@10
          - type: dot_precision@1
            value: 0.35333333333333333
            name: Dot Precision@1
          - type: dot_precision@3
            value: 0.24444444444444438
            name: Dot Precision@3
          - type: dot_precision@5
            value: 0.19600000000000004
            name: Dot Precision@5
          - type: dot_precision@10
            value: 0.13066666666666665
            name: Dot Precision@10
          - type: dot_recall@1
            value: 0.21415974018873687
            name: Dot Recall@1
          - type: dot_recall@3
            value: 0.3787075790108851
            name: Dot Recall@3
          - type: dot_recall@5
            value: 0.4608539371712423
            name: Dot Recall@5
          - type: dot_recall@10
            value: 0.5112305308948334
            name: Dot Recall@10
          - type: dot_ndcg@10
            value: 0.4457050545976819
            name: Dot Ndcg@10
          - type: dot_mrr@10
            value: 0.46443915343915343
            name: Dot Mrr@10
          - type: dot_map@100
            value: 0.3500678077198655
            name: Dot Map@100
          - type: query_active_dims
            value: 7.733333110809326
            name: Query Active Dims
          - type: query_sparsity_ratio
            value: 0.999746630852801
            name: Query Sparsity Ratio
          - type: corpus_active_dims
            value: 64.12316794619026
            name: Corpus Active Dims
          - type: corpus_sparsity_ratio
            value: 0.9978991164423633
            name: Corpus Sparsity Ratio

Inference-free SPLADE bert-base-uncased trained on Natural-Questions tuples

This is a Asymmetric Inference-free SPLADE Sparse Encoder model trained on the natural-questions dataset 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: Asymmetric Inference-free SPLADE Sparse Encoder
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 30522 dimensions
• Similarity Function: Dot Product
• Training Dataset:
  • natural-questions
• 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): Router(
    (query_0_IDF): IDF ({'frozen': False}, dim:30522, tokenizer: BertTokenizerFast)
    (document_0_MLMTransformer): MLMTransformer({'max_seq_length': 512, 'do_lower_case': False}) with MLMTransformer model: BertForMaskedLM 
    (document_1_SpladePooling): 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/inference-free-splade-bert-base-uncased-nq")
# Run inference
queries = [
    "is send in the clowns from a musical",
]
documents = [
    'Send In the Clowns "Send In the Clowns" is a song written by Stephen Sondheim for the 1973 musical A Little Night Music, an adaptation of Ingmar Bergman\'s film Smiles of a Summer Night. It is a ballad from Act Two, in which the character Desirée reflects on the ironies and disappointments of her life. Among other things, she looks back on an affair years earlier with the lawyer Fredrik, who was deeply in love with her but whose marriage proposals she had rejected. Meeting him after so long, she realizes she is in love with him and finally ready to marry him, but now it is he who rejects her: he is in an unconsummated marriage with a much younger woman. Desirée proposes marriage to rescue him from this situation, but he declines, citing his dedication to his bride. Reacting to his rejection, Desirée sings this song. The song is later reprised as a coda after Fredrik\'s young wife runs away with his son, and Fredrik is finally free to accept Desirée\'s offer.[1]',
    'The Suite Life on Deck The Suite Life on Deck is an American sitcom that aired on Disney Channel from September 26, 2008 to May 6, 2011. It is a sequel/spin-off of the Disney Channel Original Series The Suite Life of Zack & Cody. The series follows twin brothers Zack and Cody Martin and hotel heiress London Tipton in a new setting, the SS Tipton, where they attend classes at "Seven Seas High School" and meet Bailey Pickett while Mr. Moseby manages the ship. The ship travels around the world to nations such as Italy, France, Greece, India, Sweden and the United Kingdom where the characters experience different cultures, adventures, and situations.[1]',
    'Money in the Bank ladder match The first match was contested in 2005 at WrestleMania 21, after being invented (in kayfabe) by Chris Jericho.[1] At the time, it was exclusive to wrestlers of the Raw brand, and Edge won the inaugural match.[1] From then until 2010, the Money in the Bank ladder match, now open to all WWE brands, became a WrestleMania mainstay. 2010 saw a second and third Money in the Bank ladder match when the Money in the Bank pay-per-view debuted in July. Unlike the matches at WrestleMania, this new event featured two such ladder matches – one each for a contract for the WWE Championship and World Heavyweight Championship, respectively.',
]
query_embeddings = model.encode_query(queries)
document_embeddings = model.encode_documents(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([[8.1355, 0.0000, 0.1926]])

Evaluation

Metrics

Sparse Information Retrieval

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

Metric	NanoMSMARCO	NanoNFCorpus	NanoNQ
dot_accuracy@1	0.24	0.44	0.38
dot_accuracy@3	0.54	0.54	0.54
dot_accuracy@5	0.68	0.56	0.62
dot_accuracy@10	0.78	0.58	0.66
dot_precision@1	0.24	0.44	0.38
dot_precision@3	0.18	0.3733	0.18
dot_precision@5	0.136	0.324	0.128
dot_precision@10	0.078	0.246	0.068
dot_recall@1	0.24	0.0225	0.38
dot_recall@3	0.54	0.0761	0.52
dot_recall@5	0.68	0.0926	0.61
dot_recall@10	0.78	0.1137	0.64
dot_ndcg@10	0.509	0.3094	0.5188
dot_mrr@10	0.4219	0.4907	0.4807
dot_map@100	0.4321	0.1284	0.4897
query_active_dims	7.22	5.66	10.32
query_sparsity_ratio	0.9998	0.9998	0.9997
corpus_active_dims	64.5061	81.6082	53.4848
corpus_sparsity_ratio	0.9979	0.9973	0.9982

Sparse Nano BEIR

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

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

Metric	Value
dot_accuracy@1	0.3533
dot_accuracy@3	0.54
dot_accuracy@5	0.62
dot_accuracy@10	0.6733
dot_precision@1	0.3533
dot_precision@3	0.2444
dot_precision@5	0.196
dot_precision@10	0.1307
dot_recall@1	0.2142
dot_recall@3	0.3787
dot_recall@5	0.4609
dot_recall@10	0.5112
dot_ndcg@10	0.4457
dot_mrr@10	0.4644
dot_map@100	0.3501
query_active_dims	7.7333
query_sparsity_ratio	0.9997
corpus_active_dims	64.1232
corpus_sparsity_ratio	0.9979

Training Details

Training Dataset

natural-questions

• Dataset: natural-questions at f9e894e
• Size: 99,000 training samples
• Columns: query and answer
• Approximate statistics based on the first 1000 samples:

  	query	answer
  type	string	string
  details	min: 10 tokens mean: 11.71 tokens max: 26 tokens	min: 4 tokens mean: 131.81 tokens max: 450 tokens

• Samples:

  query	answer
  who played the father in papa don't preach	Alex McArthur Alex McArthur (born March 6, 1957) is an American actor.
  where was the location of the battle of hastings	Battle of Hastings The Battle of Hastings[a] was fought on 14 October 1066 between the Norman-French army of William, the Duke of Normandy, and an English army under the Anglo-Saxon King Harold Godwinson, beginning the Norman conquest of England. It took place approximately 7 miles (11 kilometres) northwest of Hastings, close to the present-day town of Battle, East Sussex, and was a decisive Norman victory.
  how many puppies can a dog give birth to	Canine reproduction The largest litter size to date was set by a Neapolitan Mastiff in Manea, Cambridgeshire, UK on November 29, 2004; the litter was 24 puppies.[22]

• Loss: SpladeLoss with these parameters:

  {
      "loss": "SparseMultipleNegativesRankingLoss(scale=1.0, similarity_fct='dot_score')",
      "lambda_corpus": 0.003,
      "lambda_query": 0
  }
  

Evaluation Dataset

natural-questions

• Dataset: natural-questions at f9e894e
• Size: 1,000 evaluation samples
• Columns: query and answer
• Approximate statistics based on the first 1000 samples:

  	query	answer
  type	string	string
  details	min: 10 tokens mean: 11.69 tokens max: 23 tokens	min: 15 tokens mean: 134.01 tokens max: 512 tokens

• Samples:

  query	answer
  where is the tiber river located in italy	Tiber The Tiber (/ˈtaɪbər/, Latin: Tiberis,[1] Italian: Tevere [ˈteːvere])[2] is the third-longest river in Italy, rising in the Apennine Mountains in Emilia-Romagna and flowing 406 kilometres (252 mi) through Tuscany, Umbria and Lazio, where it is joined by the river Aniene, to the Tyrrhenian Sea, between Ostia and Fiumicino.[3] It drains a basin estimated at 17,375 square kilometres (6,709 sq mi). The river has achieved lasting fame as the main watercourse of the city of Rome, founded on its eastern banks.
  what kind of car does jay gatsby drive	Jay Gatsby At the Buchanan home, Jordan Baker, Nick, Jay, and the Buchanans decide to visit New York City. Tom borrows Gatsby's yellow Rolls Royce to drive up to the city. On the way to New York City, Tom makes a detour at a gas station in "the Valley of Ashes", a run-down part of Long Island. The owner, George Wilson, shares his concern that his wife, Myrtle, may be having an affair. This unnerves Tom, who has been having an affair with Myrtle, and he leaves in a hurry.
  who sings if i can dream about you	I Can Dream About You "I Can Dream About You" is a song performed by American singer Dan Hartman on the soundtrack album of the film Streets of Fire. Released in 1984 as a single from the soundtrack, and included on Hartman's album I Can Dream About You, it reached number 6 on the Billboard Hot 100.[1]

• Loss: SpladeLoss with these parameters:

  {
      "loss": "SparseMultipleNegativesRankingLoss(scale=1.0, similarity_fct='dot_score')",
      "lambda_corpus": 0.003,
      "lambda_query": 0
  }
  

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
• router_mapping: {'query': 'query', 'answer': 'document'}
• learning_rate_mapping: {'IDF\.weight': 0.001}

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: {'query': 'query', 'answer': 'document'}
• learning_rate_mapping: {'IDF\.weight': 0.001}

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
0.0323	200	0.2623	-	-	-	-	-
0.0646	400	0.103	0.1146	0.5168	0.3151	0.5116	0.4479
0.0970	600	0.0867	-	-	-	-	-
0.1293	800	0.0656	0.0912	0.5291	0.3181	0.4704	0.4392
0.1616	1000	0.0668	-	-	-	-	-
0.1939	1200	0.0737	0.0962	0.4889	0.3257	0.5003	0.4383
0.2262	1400	0.08	-	-	-	-	-
0.2586	1600	0.0826	0.0943	0.5382	0.3263	0.5141	0.4595
0.2909	1800	0.0863	-	-	-	-	-
0.3232	2000	0.0896	0.1025	0.5305	0.3138	0.5585	0.4676
0.3555	2200	0.089	-	-	-	-	-
0.3878	2400	0.0845	0.1030	0.5135	0.3281	0.5277	0.4564
0.4202	2600	0.091	-	-	-	-	-
0.4525	2800	0.0837	0.0946	0.5384	0.3284	0.5282	0.4650
0.4848	3000	0.0761	-	-	-	-	-
0.5171	3200	0.0776	0.0917	0.5059	0.3253	0.5169	0.4494
0.5495	3400	0.0814	-	-	-	-	-
0.5818	3600	0.0767	0.0862	0.5381	0.3188	0.5323	0.4631
0.6141	3800	0.0824	-	-	-	-	-
0.6464	4000	0.0734	0.0853	0.5392	0.3234	0.5192	0.4606
0.6787	4200	0.0769	-	-	-	-	-
0.7111	4400	0.0681	0.0825	0.5236	0.3187	0.5369	0.4597
0.7434	4600	0.0699	-	-	-	-	-
0.7757	4800	0.0702	0.0814	0.5267	0.2961	0.5420	0.4549
0.8080	5000	0.0708	-	-	-	-	-
0.8403	5200	0.0731	0.0765	0.5376	0.3020	0.5307	0.4568
0.8727	5400	0.0739	-	-	-	-	-
0.9050	5600	0.0669	0.0771	0.5160	0.3091	0.5455	0.4569
0.9373	5800	0.0658	-	-	-	-	-
0.9696	6000	0.0662	0.0755	0.5080	0.3105	0.5215	0.4467
-1	-1	-	-	0.5090	0.3094	0.5188	0.4457

Environmental Impact

Carbon emissions were measured using CodeCarbon.

• Energy Consumed: 0.164 kWh
• Carbon Emitted: 0.064 kg of CO2
• Hours Used: 0.432 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: 2.6.0+cu124
• 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},
}

SparseMultipleNegativesRankingLoss

@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}
}

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}
    }