SparseEncoder based on microsoft/mpnet-base

This is a sentence-transformers model finetuned from microsoft/mpnet-base on the natural-questions dataset. 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.

See train_nq.py for the training script used for this model.

Warning: Sparse models in Sentence Transformers are still quite experimental.

Model Details

Model Description

Model Type: Sentence Transformer
Base model: microsoft/mpnet-base
Maximum Sequence Length: 512 tokens
Output Dimensionality: 768 dimensions
Similarity Function: Cosine Similarity
Training Dataset:
- natural-questions
Language: en

Model Sources

Documentation: Sentence Transformers Documentation
Repository: Sentence Transformers on GitHub
Hugging Face: Sentence Transformers on Hugging Face

Full Model Architecture

SparseEncoder(
  (0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: MPNetModel 
  (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})
  (2): CSRSparsity({'input_dim': 768, 'hidden_dim': 3072, 'k': 256, 'k_aux': 512, 'normalize': False, 'dead_threshold': 30})
)

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("tomaarsen/sparse-mpnet-base-nq-fresh")
# Run inference
sentences = [
    'who is cornelius in the book of acts',
    'Cornelius the Centurion Cornelius (Greek: Κορνήλιος) was a Roman centurion who is considered by Christians to be one of the first Gentiles to convert to the faith, as related in Acts of the Apostles.',
    "Joe Ranft Ranft reunited with Lasseter when he was hired by Pixar in 1991 as their head of story.[1] There he worked on all of their films produced up to 2006; this included Toy Story (for which he received an Academy Award nomination) and A Bug's Life, as the co-story writer and others as story supervisor. His final film was Cars. He also voiced characters in many of the films, including Heimlich the caterpillar in A Bug's Life, Wheezy the penguin in Toy Story 2, and Jacques the shrimp in Finding Nemo.[1]",
]
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

Sparse Information Retrieval

Datasets: NanoMSMARCO_16, NanoNFCorpus_16 and NanoNQ_16
Evaluated with SparseInformationRetrievalEvaluator with these parameters:
```
{
    "truncate_dim": 16
}
```

Metric	NanoMSMARCO_16	NanoNFCorpus_16	NanoNQ_16
cosine_accuracy@1	0.1	0.08	0.18
cosine_accuracy@3	0.26	0.14	0.42
cosine_accuracy@5	0.36	0.24	0.54
cosine_accuracy@10	0.5	0.32	0.64
cosine_precision@1	0.1	0.08	0.18
cosine_precision@3	0.0867	0.06	0.14
cosine_precision@5	0.072	0.08	0.108
cosine_precision@10	0.05	0.05	0.064
cosine_recall@1	0.1	0.006	0.18
cosine_recall@3	0.26	0.0094	0.4
cosine_recall@5	0.36	0.0133	0.5
cosine_recall@10	0.5	0.0165	0.6
cosine_ndcg@10	0.2721	0.061	0.3867
cosine_mrr@10	0.2023	0.1407	0.3267
cosine_map@100	0.2176	0.0153	0.325

Sparse Nano BEIR

Dataset: NanoBEIR_mean_16

Evaluated with SparseNanoBEIREvaluator with these parameters:

{
    "dataset_names": [
        "msmarco",
        "nfcorpus",
        "nq"
    ],
    "truncate_dim": 16
}

Metric	Value
cosine_accuracy@1	0.12
cosine_accuracy@3	0.2733
cosine_accuracy@5	0.38
cosine_accuracy@10	0.4867
cosine_precision@1	0.12
cosine_precision@3	0.0956
cosine_precision@5	0.0867
cosine_precision@10	0.0547
cosine_recall@1	0.0953
cosine_recall@3	0.2231
cosine_recall@5	0.2911
cosine_recall@10	0.3722
cosine_ndcg@10	0.2399
cosine_mrr@10	0.2233
cosine_map@100	0.186

Sparse Information Retrieval

Datasets: NanoMSMARCO_32, NanoNFCorpus_32 and NanoNQ_32
Evaluated with SparseInformationRetrievalEvaluator with these parameters:
```
{
    "truncate_dim": 32
}
```

Metric	NanoMSMARCO_32	NanoNFCorpus_32	NanoNQ_32
cosine_accuracy@1	0.18	0.14	0.32
cosine_accuracy@3	0.26	0.26	0.46
cosine_accuracy@5	0.36	0.28	0.58
cosine_accuracy@10	0.56	0.34	0.68
cosine_precision@1	0.18	0.14	0.32
cosine_precision@3	0.0867	0.1133	0.1533
cosine_precision@5	0.072	0.096	0.116
cosine_precision@10	0.056	0.09	0.068
cosine_recall@1	0.18	0.0077	0.31
cosine_recall@3	0.26	0.0123	0.42
cosine_recall@5	0.36	0.017	0.53
cosine_recall@10	0.56	0.0242	0.63
cosine_ndcg@10	0.3311	0.1023	0.4604
cosine_mrr@10	0.2634	0.2055	0.4212
cosine_map@100	0.2794	0.0226	0.4113

Sparse Nano BEIR

Dataset: NanoBEIR_mean_32

Evaluated with SparseNanoBEIREvaluator with these parameters:

{
    "dataset_names": [
        "msmarco",
        "nfcorpus",
        "nq"
    ],
    "truncate_dim": 32
}

Metric	Value
cosine_accuracy@1	0.2133
cosine_accuracy@3	0.3267
cosine_accuracy@5	0.4067
cosine_accuracy@10	0.5267
cosine_precision@1	0.2133
cosine_precision@3	0.1178
cosine_precision@5	0.0947
cosine_precision@10	0.0713
cosine_recall@1	0.1659
cosine_recall@3	0.2308
cosine_recall@5	0.3023
cosine_recall@10	0.4047
cosine_ndcg@10	0.2979
cosine_mrr@10	0.2967
cosine_map@100	0.2377

Sparse Information Retrieval

Datasets: NanoMSMARCO_64, NanoNFCorpus_64 and NanoNQ_64
Evaluated with SparseInformationRetrievalEvaluator with these parameters:
```
{
    "truncate_dim": 64
}
```

Metric	NanoMSMARCO_64	NanoNFCorpus_64	NanoNQ_64
cosine_accuracy@1	0.16	0.18	0.44
cosine_accuracy@3	0.38	0.26	0.62
cosine_accuracy@5	0.46	0.32	0.68
cosine_accuracy@10	0.6	0.4	0.72
cosine_precision@1	0.16	0.18	0.44
cosine_precision@3	0.1267	0.1267	0.2067
cosine_precision@5	0.092	0.12	0.14
cosine_precision@10	0.06	0.088	0.074
cosine_recall@1	0.16	0.0095	0.42
cosine_recall@3	0.38	0.0129	0.58
cosine_recall@5	0.46	0.0369	0.64
cosine_recall@10	0.6	0.0476	0.68
cosine_ndcg@10	0.3545	0.115	0.5619
cosine_mrr@10	0.278	0.2421	0.5396
cosine_map@100	0.2957	0.0318	0.5268

Sparse Nano BEIR

Dataset: NanoBEIR_mean_64

Evaluated with SparseNanoBEIREvaluator with these parameters:

{
    "dataset_names": [
        "msmarco",
        "nfcorpus",
        "nq"
    ],
    "truncate_dim": 64
}

Metric	Value
cosine_accuracy@1	0.26
cosine_accuracy@3	0.42
cosine_accuracy@5	0.4867
cosine_accuracy@10	0.5733
cosine_precision@1	0.26
cosine_precision@3	0.1533
cosine_precision@5	0.1173
cosine_precision@10	0.074
cosine_recall@1	0.1965
cosine_recall@3	0.3243
cosine_recall@5	0.379
cosine_recall@10	0.4425
cosine_ndcg@10	0.3438
cosine_mrr@10	0.3532
cosine_map@100	0.2848

Sparse Information Retrieval

Datasets: NanoMSMARCO_128, NanoNFCorpus_128 and NanoNQ_128
Evaluated with SparseInformationRetrievalEvaluator with these parameters:
```
{
    "truncate_dim": 128
}
```

Metric	NanoMSMARCO_128	NanoNFCorpus_128	NanoNQ_128
cosine_accuracy@1	0.2	0.14	0.38
cosine_accuracy@3	0.34	0.34	0.56
cosine_accuracy@5	0.46	0.38	0.7
cosine_accuracy@10	0.68	0.52	0.8
cosine_precision@1	0.2	0.14	0.38
cosine_precision@3	0.1133	0.1667	0.1867
cosine_precision@5	0.092	0.128	0.144
cosine_precision@10	0.068	0.114	0.082
cosine_recall@1	0.2	0.0037	0.35
cosine_recall@3	0.34	0.0212	0.53
cosine_recall@5	0.46	0.0246	0.66
cosine_recall@10	0.68	0.0433	0.76
cosine_ndcg@10	0.4022	0.1267	0.5527
cosine_mrr@10	0.3182	0.2538	0.5072
cosine_map@100	0.3323	0.0333	0.4847

Sparse Nano BEIR

Dataset: NanoBEIR_mean_128

Evaluated with SparseNanoBEIREvaluator with these parameters:

{
    "dataset_names": [
        "msmarco",
        "nfcorpus",
        "nq"
    ],
    "truncate_dim": 128
}

Metric	Value
cosine_accuracy@1	0.24
cosine_accuracy@3	0.4133
cosine_accuracy@5	0.5133
cosine_accuracy@10	0.6667
cosine_precision@1	0.24
cosine_precision@3	0.1556
cosine_precision@5	0.1213
cosine_precision@10	0.088
cosine_recall@1	0.1846
cosine_recall@3	0.2971
cosine_recall@5	0.3815
cosine_recall@10	0.4944
cosine_ndcg@10	0.3605
cosine_mrr@10	0.3597
cosine_map@100	0.2834

Sparse Information Retrieval

Datasets: NanoMSMARCO_256, NanoNFCorpus_256 and NanoNQ_256
Evaluated with SparseInformationRetrievalEvaluator with these parameters:
```
{
    "truncate_dim": 256
}
```

Metric	NanoMSMARCO_256	NanoNFCorpus_256	NanoNQ_256
cosine_accuracy@1	0.26	0.18	0.42
cosine_accuracy@3	0.48	0.28	0.58
cosine_accuracy@5	0.52	0.38	0.68
cosine_accuracy@10	0.68	0.5	0.76
cosine_precision@1	0.26	0.18	0.42
cosine_precision@3	0.16	0.1467	0.1933
cosine_precision@5	0.104	0.14	0.14
cosine_precision@10	0.068	0.114	0.08
cosine_recall@1	0.26	0.0055	0.4
cosine_recall@3	0.48	0.0114	0.54
cosine_recall@5	0.52	0.0213	0.64
cosine_recall@10	0.68	0.0347	0.73
cosine_ndcg@10	0.4652	0.1263	0.5612
cosine_mrr@10	0.398	0.2575	0.5227
cosine_map@100	0.4125	0.0337	0.5087

Sparse Nano BEIR

Dataset: NanoBEIR_mean_256

Evaluated with SparseNanoBEIREvaluator with these parameters:

{
    "dataset_names": [
        "msmarco",
        "nfcorpus",
        "nq"
    ],
    "truncate_dim": 256
}

Metric	Value
cosine_accuracy@1	0.2867
cosine_accuracy@3	0.4467
cosine_accuracy@5	0.5267
cosine_accuracy@10	0.6467
cosine_precision@1	0.2867
cosine_precision@3	0.1667
cosine_precision@5	0.128
cosine_precision@10	0.0873
cosine_recall@1	0.2218
cosine_recall@3	0.3438
cosine_recall@5	0.3938
cosine_recall@10	0.4816
cosine_ndcg@10	0.3842
cosine_mrr@10	0.3927
cosine_map@100	0.3183

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

	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: CSRLoss with these parameters:

{
    "beta": 0.1,
    "gamma": 1,
    "scale": 20.0
}

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

	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: CSRLoss with these parameters:

{
    "beta": 0.1,
    "gamma": 1,
    "scale": 20.0
}

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: steps
per_device_train_batch_size: 32
per_device_eval_batch_size: 32
learning_rate: 4e-05
weight_decay: 0.0001
adam_epsilon: 6.25e-10
num_train_epochs: 1
warmup_ratio: 0.1
bf16: 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: 32
per_device_eval_batch_size: 32
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: 4e-05
weight_decay: 0.0001
adam_beta1: 0.9
adam_beta2: 0.999
adam_epsilon: 6.25e-10
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: True
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: 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}
tp_size: 0
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

Training Logs

Epoch	Step	Training Loss	Validation Loss	NanoMSMARCO_16_cosine_ndcg@10	NanoNFCorpus_16_cosine_ndcg@10	NanoNQ_16_cosine_ndcg@10	NanoBEIR_mean_16_cosine_ndcg@10	NanoMSMARCO_32_cosine_ndcg@10	NanoNFCorpus_32_cosine_ndcg@10	NanoNQ_32_cosine_ndcg@10	NanoBEIR_mean_32_cosine_ndcg@10	NanoMSMARCO_64_cosine_ndcg@10	NanoNFCorpus_64_cosine_ndcg@10	NanoNQ_64_cosine_ndcg@10	NanoBEIR_mean_64_cosine_ndcg@10	NanoMSMARCO_128_cosine_ndcg@10	NanoNFCorpus_128_cosine_ndcg@10	NanoNQ_128_cosine_ndcg@10	NanoBEIR_mean_128_cosine_ndcg@10	NanoMSMARCO_256_cosine_ndcg@10	NanoNFCorpus_256_cosine_ndcg@10	NanoNQ_256_cosine_ndcg@10	NanoBEIR_mean_256_cosine_ndcg@10
-1	-1	-	-	0.0318	0.0148	0.0149	0.0205	0.0794	0.0234	0.0102	0.0377	0.0855	0.0195	0.0508	0.0519	0.1081	0.0246	0.0264	0.0530	0.1006	0.0249	0.0388	0.0547
0.0646	200	0.7332	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
0.1293	400	0.2606	0.1970	0.2845	0.0970	0.3546	0.2454	0.3778	0.1358	0.3455	0.2864	0.3868	0.1563	0.3806	0.3079	0.3988	0.1664	0.4035	0.3229	0.4020	0.1782	0.4181	0.3327
0.1939	600	0.2247	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
0.2586	800	0.1983	0.1750	0.2908	0.0866	0.3730	0.2502	0.3324	0.1155	0.4275	0.2918	0.3511	0.1621	0.4998	0.3377	0.3920	0.1563	0.5174	0.3553	0.4152	0.1555	0.5153	0.3620
0.3232	1000	0.1822	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
0.3878	1200	0.1846	0.1594	0.2775	0.0785	0.3723	0.2428	0.2642	0.1076	0.4389	0.2702	0.3865	0.1328	0.4329	0.3174	0.3883	0.1446	0.5040	0.3456	0.3638	0.1529	0.4939	0.3369
0.4525	1400	0.1669	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
0.5171	1600	0.1573	0.1452	0.2740	0.0624	0.3670	0.2345	0.3557	0.0855	0.4188	0.2867	0.4094	0.1099	0.5027	0.3407	0.3885	0.1340	0.4990	0.3405	0.4820	0.1577	0.5453	0.3950
0.5818	1800	0.1502	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
0.6464	2000	0.1375	0.1255	0.2307	0.0685	0.3801	0.2264	0.2529	0.0815	0.4335	0.2560	0.3509	0.0955	0.4611	0.3025	0.3932	0.1339	0.4875	0.3382	0.4184	0.1483	0.4904	0.3523
0.7111	2200	0.1359	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
0.7757	2400	0.1288	0.1184	0.2737	0.0703	0.3419	0.2286	0.3765	0.0843	0.4440	0.3016	0.3927	0.1247	0.5285	0.3486	0.3726	0.1203	0.5153	0.3361	0.4676	0.1343	0.5523	0.3847
0.8403	2600	0.1235	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
0.9050	2800	0.1168	0.1094	0.2751	0.0710	0.3602	0.2354	0.3227	0.0966	0.5046	0.3080	0.4112	0.1129	0.5268	0.3503	0.4077	0.1259	0.5253	0.3530	0.4642	0.1238	0.5726	0.3869
0.9696	3000	0.1187	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
-1	-1	-	-	0.2721	0.0610	0.3867	0.2399	0.3311	0.1023	0.4604	0.2979	0.3545	0.1150	0.5619	0.3438	0.4022	0.1267	0.5527	0.3605	0.4652	0.1263	0.5612	0.3842

Environmental Impact

Carbon emissions were measured using CodeCarbon.

Energy Consumed: 0.292 kWh
Carbon Emitted: 0.113 kg of CO2
Hours Used: 0.773 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: 4.1.0.dev0
Transformers: 4.52.0.dev0
PyTorch: 2.6.0+cu124
Accelerate: 1.5.1
Datasets: 3.3.2
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",
}

tomaarsen
/

sparse-mpnet-base-nq-fresh

SparseEncoder based on microsoft/mpnet-base

Model Details

Model Description

Model Sources

Full Model Architecture

Usage

Direct Usage (Sentence Transformers)

Evaluation

Metrics

Sparse Information Retrieval

Sparse Nano BEIR

Sparse Information Retrieval

Sparse Nano BEIR

Sparse Information Retrieval

Sparse Nano BEIR

Sparse Information Retrieval

Sparse Nano BEIR

Sparse Information Retrieval

Sparse Nano BEIR

Training Details

Training Dataset

natural-questions

Evaluation Dataset

natural-questions

Training Hyperparameters

Non-Default Hyperparameters

All Hyperparameters

Training Logs

Environmental Impact

Training Hardware

Framework Versions

Citation

BibTeX

Sentence Transformers

Model tree for tomaarsen/sparse-mpnet-base-nq-fresh

Dataset used to train tomaarsen/sparse-mpnet-base-nq-fresh

Evaluation results