MiniLM-L12-H384 trained on GooAQ

This is a Cross Encoder model finetuned from microsoft/MiniLM-L12-H384-uncased using the sentence-transformers library. It computes scores for pairs of texts, which can be used for text reranking and semantic search.

Model Details

Model Description

• Model Type: Cross Encoder
• Base model: microsoft/MiniLM-L12-H384-uncased
• Maximum Sequence Length: 512 tokens
• Number of Output Labels: 1 label

• Language: en
• License: apache-2.0

Model Sources

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

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 CrossEncoder

# Download from the 🤗 Hub
model = CrossEncoder("zhensuuu/reranker-MiniLM-L12-H384-uncased-intent")
# Get scores for pairs of texts
pairs = [
    ['Add edge representing resource request', ' Model process-resource dependency relationship'],
    ['Split text into words list', ' Filter words matching given keyword.'],
    ['Calculate approximate cube root value', ' Find cube root using exponentiation'],
    ['Reverse sublist within linked list', ' Move nodes to new positions'],
    ['Defines neighbors for node A', ' Specifies direct connections from A'],
]
scores = model.predict(pairs)
print(scores.shape)
# (5,)

# Or rank different texts based on similarity to a single text
ranks = model.rank(
    'Add edge representing resource request',
    [
        ' Model process-resource dependency relationship',
        ' Filter words matching given keyword.',
        ' Find cube root using exponentiation',
        ' Move nodes to new positions',
        ' Specifies direct connections from A',
    ]
)
# [{'corpus_id': ..., 'score': ...}, {'corpus_id': ..., 'score': ...}, ...]

Evaluation

Metrics

Cross Encoder Reranking

• Datasets: NanoMSMARCO_R100, NanoNFCorpus_R100 and NanoNQ_R100
• Evaluated with CrossEncoderRerankingEvaluator with these parameters:

  {
      "at_k": 10,
      "always_rerank_positives": true
  }
  

Cross Encoder Nano BEIR

• Dataset: NanoBEIR_R100_mean
• Evaluated with CrossEncoderNanoBEIREvaluator with these parameters:

  {
      "dataset_names": [
          "msmarco",
          "nfcorpus",
          "nq"
      ],
      "rerank_k": 100,
      "at_k": 10,
      "always_rerank_positives": true
  }
  

Training Details

Training Dataset

Unnamed Dataset

• Size: 85,938 training samples
• Columns: question and answer
• Approximate statistics based on the first 1000 samples:

  	question	answer
  type	string	string
  details	min: 18 characters mean: 33.49 characters max: 49 characters	min: 18 characters mean: 35.88 characters max: 52 characters

• Samples:

  question	answer
  Check if configuration loaded successfully	prevent further actions if configuration absent
  Add new user to list	Store received user in memory
  Selects profitable jobs and schedules	Displays scheduled jobs and profit

• Loss: CachedMultipleNegativesRankingLoss with these parameters:

  {
      "scale": 10.0,
      "num_negatives": 5,
      "activation_fn": "torch.nn.modules.activation.Sigmoid",
      "mini_batch_size": 16
  }
  

Evaluation Dataset

Unnamed Dataset

• Size: 1,000 evaluation samples
• Columns: question and answer
• Approximate statistics based on the first 1000 samples:

  	question	answer
  type	string	string
  details	min: 20 characters mean: 33.63 characters max: 54 characters	min: 18 characters mean: 35.86 characters max: 55 characters

• Samples:

  question	answer
  Add edge representing resource request	Model process-resource dependency relationship
  Split text into words list	Filter words matching given keyword.
  Calculate approximate cube root value	Find cube root using exponentiation

• Loss: CachedMultipleNegativesRankingLoss with these parameters:

  {
      "scale": 10.0,
      "num_negatives": 5,
      "activation_fn": "torch.nn.modules.activation.Sigmoid",
      "mini_batch_size": 16
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 64
• per_device_eval_batch_size: 64
• learning_rate: 2e-05
• num_train_epochs: 1
• warmup_ratio: 0.1
• seed: 12
• bf16: 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: 64
• per_device_eval_batch_size: 64
• 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: 12
• 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}
• 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
• dispatch_batches: None
• split_batches: 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: batch_sampler
• multi_dataset_batch_sampler: proportional
• router_mapping: {}
• learning_rate_mapping: {}

Training Logs

Environmental Impact

Carbon emissions were measured using CodeCarbon.

• Energy Consumed: 0.082 kWh
• Carbon Emitted: 0.000 kg of CO2
• Hours Used: 0.306 hours

Training Hardware

• On Cloud: No
• GPU Model: 4 x NVIDIA RTX 6000 Ada Generation
• CPU Model: AMD EPYC 7763 64-Core Processor
• RAM Size: 251.53 GB

Framework Versions

• Python: 3.10.12
• Sentence Transformers: 5.1.0
• Transformers: 4.48.3
• PyTorch: 2.6.0+cu124
• Accelerate: 1.6.0
• Datasets: 3.6.0
• 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",
}