redis/model-a-baseline

SentenceTransformer based on thenlper/gte-small

This is a sentence-transformers model finetuned from thenlper/gte-small. It maps sentences & paragraphs to a 384-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.

Model Details

Model Description

• Model Type: Sentence Transformer
• Base model: thenlper/gte-small
• Maximum Sequence Length: 128 tokens
• Output Dimensionality: 384 dimensions
• 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': 128, 'do_lower_case': False, 'architecture': 'BertModel'})
  (1): Pooling({'word_embedding_dimension': 384, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
  (2): Normalize()
)

Usage

Direct Usage (Sentence Transformers)

First install the Sentence Transformers library:

pip install -U sentence-transformers

Then you can load this model and run inference.

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("redis/model-a-baseline")
# Run inference
sentences = [
    'Why do onions make people cry?',
    'Why do onions sting?',
    'Can people with bipolar have healthy relationships?',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 384]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities)
# tensor([[1.0000, 0.7584, 0.1067],
#         [0.7584, 1.0000, 0.0503],
#         [0.1067, 0.0503, 1.0000]])

Evaluation

Metrics

Information Retrieval

• Datasets: NanoMSMARCO and NanoNQ
• Evaluated with InformationRetrievalEvaluator

Metric	NanoMSMARCO	NanoNQ
cosine_accuracy@1	0.3	0.38
cosine_accuracy@3	0.62	0.56
cosine_accuracy@5	0.7	0.62
cosine_accuracy@10	0.76	0.7
cosine_precision@1	0.3	0.38
cosine_precision@3	0.2067	0.1933
cosine_precision@5	0.14	0.132
cosine_precision@10	0.076	0.074
cosine_recall@1	0.3	0.35
cosine_recall@3	0.62	0.54
cosine_recall@5	0.7	0.6
cosine_recall@10	0.76	0.68
cosine_ndcg@10	0.5416	0.5219
cosine_mrr@10	0.4702	0.4853
cosine_map@100	0.4817	0.4742

Nano BEIR

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

  {
      "dataset_names": [
          "msmarco",
          "nq"
      ],
      "dataset_id": "lightonai/NanoBEIR-en"
  }
  

Metric	Value
cosine_accuracy@1	0.34
cosine_accuracy@3	0.59
cosine_accuracy@5	0.66
cosine_accuracy@10	0.73
cosine_precision@1	0.34
cosine_precision@3	0.2
cosine_precision@5	0.136
cosine_precision@10	0.075
cosine_recall@1	0.325
cosine_recall@3	0.58
cosine_recall@5	0.65
cosine_recall@10	0.72
cosine_ndcg@10	0.5318
cosine_mrr@10	0.4778
cosine_map@100	0.478

Training Details

Training Dataset

Unnamed Dataset

• Size: 89,998 training samples
• Columns: anchor, positive, and negative
• Approximate statistics based on the first 1000 samples:

  	anchor	positive	negative
  type	string	string	string
  details	min: 5 tokens mean: 15.61 tokens max: 71 tokens	min: 5 tokens mean: 15.72 tokens max: 71 tokens	min: 4 tokens mean: 16.55 tokens max: 67 tokens

• Samples:

  anchor	positive	negative
  How long did it take to develop Pokémon GO?	How long did it take to develop Pokémon GO?	Can I take more than one gym in Pokémon GO?
  What is the best gift you've received?	What is the best tangible gift you've ever received?	Where can I download Chaayam Poosiya Veedu (The Painted House) malayalam movie for free?
  Why should I bother writing/editing a Wikipedia article when it can be overwritten by anyone?	Why should I bother writing/editing a Wikipedia article when it can be overwritten by anyone?	When I write a chapter, after I finish editing it, it is way too short. How can I lengthen a chapter?

• Loss: MultipleNegativesRankingLoss with these parameters:

  {
      "scale": 7.0,
      "similarity_fct": "cos_sim",
      "gather_across_devices": false
  }
  

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: 3 tokens mean: 15.75 tokens max: 65 tokens	min: 3 tokens mean: 15.86 tokens max: 65 tokens	min: 5 tokens mean: 16.66 tokens max: 74 tokens

• Samples:

  anchor	positive	negative
  What's it like working in IT for Goldman Sachs?	What's it like working in IT for Goldman Sachs?	What is the work done at Goldman Sachs?
  How did Revan build his foundation of his army in Star Wars?	How did Revan build his foundation of his army in Star Wars?	What Star Wars character deserves his/her own movie?
  Is C++ the best programming language to learn first?	Is C++ the best programming language to learn first?	Which programming language is the best to learn first?

• Loss: MultipleNegativesRankingLoss with these parameters:

  {
      "scale": 7.0,
      "similarity_fct": "cos_sim",
      "gather_across_devices": false
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 128
• per_device_eval_batch_size: 128
• learning_rate: 2e-05
• weight_decay: 0.0001
• max_steps: 3000
• warmup_ratio: 0.1
• fp16: True
• dataloader_drop_last: True
• dataloader_num_workers: 1
• dataloader_prefetch_factor: 1
• load_best_model_at_end: True
• optim: adamw_torch
• ddp_find_unused_parameters: False
• push_to_hub: True
• hub_model_id: redis/model-a-baseline
• eval_on_start: True

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: steps
• prediction_loss_only: True
• per_device_train_batch_size: 128
• per_device_eval_batch_size: 128
• 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: 0.0001
• adam_beta1: 0.9
• adam_beta2: 0.999
• adam_epsilon: 1e-08
• max_grad_norm: 1.0
• num_train_epochs: 3.0
• max_steps: 3000
• 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
• 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: True
• dataloader_num_workers: 1
• dataloader_prefetch_factor: 1
• 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, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
• parallelism_config: None
• deepspeed: None
• label_smoothing_factor: 0.0
• optim: adamw_torch
• optim_args: None
• adafactor: False
• group_by_length: False
• length_column_name: length
• project: huggingface
• trackio_space_id: trackio
• ddp_find_unused_parameters: False
• 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: True
• resume_from_checkpoint: None
• hub_model_id: redis/model-a-baseline
• hub_strategy: every_save
• hub_private_repo: None
• hub_always_push: False
• hub_revision: None
• 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: no
• neftune_noise_alpha: None
• optim_target_modules: None
• batch_eval_metrics: False
• eval_on_start: True
• use_liger_kernel: False
• liger_kernel_config: None
• eval_use_gather_object: False
• average_tokens_across_devices: True
• prompts: None
• batch_sampler: batch_sampler
• multi_dataset_batch_sampler: proportional
• router_mapping: {}
• learning_rate_mapping: {}

Training Logs

Epoch	Step	Training Loss	Validation Loss	NanoMSMARCO_cosine_ndcg@10	NanoNQ_cosine_ndcg@10	NanoBEIR_mean_cosine_ndcg@10
0	0	-	3.6614	0.6259	0.6583	0.6421
0.3556	250	2.1172	0.4652	0.5662	0.5558	0.5610
0.7112	500	0.5877	0.4272	0.5926	0.5289	0.5607
1.0669	750	0.5436	0.4134	0.5476	0.5478	0.5477
1.4225	1000	0.5214	0.4041	0.5461	0.5451	0.5456
1.7781	1250	0.5119	0.3994	0.5316	0.5377	0.5347
2.1337	1500	0.5006	0.3957	0.5357	0.5305	0.5331
2.4893	1750	0.494	0.3934	0.5444	0.5319	0.5382
2.8450	2000	0.4886	0.3906	0.5322	0.5246	0.5284
3.2006	2250	0.4839	0.3889	0.5413	0.5307	0.5360
3.5562	2500	0.4795	0.3881	0.5490	0.5214	0.5352
3.9118	2750	0.479	0.3872	0.5413	0.5219	0.5316
4.2674	3000	0.4765	0.3869	0.5416	0.5219	0.5318

Framework Versions

• Python: 3.10.18
• Sentence Transformers: 5.2.0
• Transformers: 4.57.3
• PyTorch: 2.9.1+cu128
• Accelerate: 1.12.0
• Datasets: 2.21.0
• Tokenizers: 0.22.1

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",
}

MultipleNegativesRankingLoss

@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}
}