iamleonie/leonies-test

SentenceTransformer based on BAAI/bge-base-en-v1.5

This is a sentence-transformers model finetuned from BAAI/bge-base-en-v1.5 on the stock_trading_qa 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.

Model Details

Model Description

• Model Type: Sentence Transformer
• Base model: BAAI/bge-base-en-v1.5
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 768 dimensions
• Similarity Function: Cosine Similarity
• Training Dataset:
  • stock_trading_qa
• Language: en

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': 512, 'do_lower_case': True}) with Transformer model: BertModel 
  (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, '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("iamleonie/leonies-test")
# Run inference
sentences = [
    'What role does back-testing play in refining event-driven trading strategies using historical data and real-time analysis?',
    'Back-testing allows traders to evaluate the performance of event-driven trading strategies using historical data, identify patterns, optimize parameters, and refine strategies for real-time implementation.',
    'Risk management techniques such as position sizing, portfolio diversification, and stop-loss orders are often used in quantitative momentum strategies to manage downside risk and protect against large losses.',
]
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

Information Retrieval

• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@3	0.675
cosine_precision@3	0.225
cosine_recall@3	0.675
cosine_ndcg@3	0.5838
cosine_mrr@3	0.5523
cosine_map@3	0.5523

Training Details

Training Dataset

stock_trading_qa

• Dataset: stock_trading_qa at 35dab2e
• Size: 6,448 training samples
• Columns: anchor and context
• Approximate statistics based on the first 1000 samples:

  	anchor	context
  type	string	string
  details	min: 7 tokens mean: 15.83 tokens max: 39 tokens	min: 17 tokens mean: 34.67 tokens max: 59 tokens

• Samples:

  anchor	context
  How should I approach investing in a volatile stock market?	Diversify your portfolio, invest in stable companies, consider dollar-cost averaging, and stay informed about market trends to make informed trading decisions.
  What is the role of cross-validation in assessing the performance of time series forecasting models for stock market trends?	Cross-validation helps evaluate the generalization ability of forecasting models by partitioning historical data into training and validation sets, ensuring that the model's performance is robust and reliable for future predictions.
  What role does correlation play in statistical arbitrage and pair trading?	Correlation measures the relationship between asset prices and helps traders identify pairs that exhibit a stable price relationship suitable for pair trading.

• Loss: MultipleNegativesRankingLoss with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim"
  }
  

Evaluation Dataset

stock_trading_qa

• Dataset: stock_trading_qa at 35dab2e
• Size: 717 evaluation samples
• Columns: anchor and context
• Approximate statistics based on the first 717 samples:

  	anchor	context
  type	string	string
  details	min: 7 tokens mean: 15.96 tokens max: 30 tokens	min: 17 tokens mean: 35.03 tokens max: 62 tokens

• Samples:

  anchor	context
  How can anomaly detection in stock prices be used to identify market inefficiencies and opportunities for arbitrage?	Anomaly detection can help identify market inefficiencies by spotting mispricings and opportunities for arbitrage, where traders can exploit price differentials to make profits by trading on anomalies.
  How do traders interpret the Head and Shoulders pattern as a trading signal?	The Head and Shoulders pattern is a reversal pattern with three peaks, where the middle peak (head) is higher than the other two (shoulders), signaling a potential trend reversal and offering a bearish trading signal.
  How do traders use Fibonacci levels as trading signals?	Fibonacci levels are used as trading signals to identify potential support and resistance levels, trend reversals, and price targets in financial markets.

• Loss: MultipleNegativesRankingLoss with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim"
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 16
• per_device_eval_batch_size: 16
• gradient_accumulation_steps: 16
• learning_rate: 2e-05
• num_train_epochs: 4
• lr_scheduler_type: cosine
• warmup_ratio: 0.1
• fp16: True
• optim: adamw_8bit
• 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: 16
• 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: 4
• max_steps: -1
• lr_scheduler_type: cosine
• 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_8bit
• 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	cosine_ndcg@3
-1	-1	-	-	0.4451
0.3970	10	5.7817	0.0765	0.5278
0.7940	20	1.295	0.0251	0.5608
1.1588	30	0.6208	0.0209	0.5771
1.5558	40	0.5701	0.0183	0.5789
1.9529	50	0.4546	0.0171	0.5882
2.3176	60	0.2861	0.0160	0.5839
2.7146	70	0.3315	0.0154	0.5818
3.0794	80	0.3179	0.0152	0.5852
3.4764	90	0.367	0.0150	0.5843
3.8734	100	0.354	0.0150	0.5838

Framework Versions

• Python: 3.11.12
• Sentence Transformers: 4.1.0
• Transformers: 4.52.4
• PyTorch: 2.6.0+cu124
• Accelerate: 1.7.0
• 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",
}

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}
}