Shashwat13333/msmarco-distilbert-base-v4_1

BGE base Financial Matryoshka

This is a sentence-transformers model finetuned from sentence-transformers/msmarco-distilbert-base-v4. 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: sentence-transformers/msmarco-distilbert-base-v4
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 768 dimensions
• Similarity Function: Cosine Similarity
• Language: en
• License: apache-2.0

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': False}) with Transformer model: DistilBertModel 
  (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})
)

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("Shashwat13333/msmarco-distilbert-base-v4")
# Run inference
sentences = [
    "What steps do you take to understand a business's needs?",
    'How do you customize your DevOps solutions for different industries?\nWe understand that each industry has unique challenges and requirements. Our approach involves a thorough analysis of your business needs, industry standards, and regulatory requirements to tailor a DevOps solution that meets your specific objectives',
    'Our Vision Be a partner for industry verticals on the inevitable journey towards enterprise transformation and future readiness, by harnessing the growing power of Artificial Intelligence, Machine Learning, Data Science and emerging methodologies, with immediacy of impact and swiftness of outcome.Our Mission\nTo decode data, and code new intelligence into products and automation, engineer, develop and deploy systems and applications that redefine experiences and realign business growth.',
]
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

• Datasets: dim_768, dim_512, dim_256, dim_128 and dim_64
• Evaluated with InformationRetrievalEvaluator

Metric	dim_768	dim_512	dim_256	dim_128	dim_64
cosine_accuracy@1	0.039	0.026	0.0519	0.0519	0.0649
cosine_accuracy@3	0.4805	0.4935	0.4675	0.4416	0.4156
cosine_accuracy@5	0.5714	0.5844	0.5195	0.5584	0.5065
cosine_accuracy@10	0.6494	0.6494	0.6234	0.6623	0.5974
cosine_precision@1	0.039	0.026	0.0519	0.0519	0.0649
cosine_precision@3	0.1602	0.1645	0.1558	0.1472	0.1385
cosine_precision@5	0.1143	0.1169	0.1039	0.1117	0.1013
cosine_precision@10	0.0649	0.0649	0.0623	0.0662	0.0597
cosine_recall@1	0.039	0.026	0.0519	0.0519	0.0649
cosine_recall@3	0.4805	0.4935	0.4675	0.4416	0.4156
cosine_recall@5	0.5714	0.5844	0.5195	0.5584	0.5065
cosine_recall@10	0.6494	0.6494	0.6234	0.6623	0.5974
cosine_ndcg@10	0.3349	0.3382	0.338	0.3429	0.3229
cosine_mrr@10	0.2338	0.237	0.2458	0.2407	0.2348
cosine_map@100	0.2465	0.2486	0.2597	0.2508	0.2482

Training Details

Training Dataset

Unnamed Dataset

• Size: 154 training samples
• Columns: anchor and positive
• Approximate statistics based on the first 154 samples:

  	anchor	positive
  type	string	string
  details	min: 7 tokens mean: 12.43 tokens max: 20 tokens	min: 20 tokens mean: 126.6 tokens max: 378 tokens

• Samples:

  anchor	positive
  What kind of websites can you help us with?	CLIENT TESTIMONIALS Worked with TCZ on two business critical website development projects. The TCZ team is a group of experts in their respective domains and have helped us with excellent end-to-end development of a website right from the conceptualization to implementation and maintenance. By Dr. Kunal Joshi - Healthcare Marketing & Strategy Professional TCZ helped us with our new website launch in a seamless manner. Through all our discussions, they made sure to have the website designed as we had envisioned it to be. Thank you team TCZ. By Dr. Sarita Ahlawat - Managing Director and Co-Founder, Botlab Dynamics
  What does DevSecOps mean?	How do you ensure the security of our DevOps pipeline? Security is a top priority in our DevOps solutions. We implement DevSecOps practices, integrating security measures into the CI/CD pipeline from the outset. This includes automated security scans, compliance checks, and vulnerability assessments to ensure your infrastructure is secure
  do you work with tech like nlp ?	What AI solutions does Techchefz specialize in? We specialize in a range of AI solutions including recommendation engines, NLP, computer vision, customer segmentation, predictive analytics, operational efficiency through machine learning, risk management, and conversational AI for customer service.

• Loss: MatryoshkaLoss with these parameters:

  {
      "loss": "MultipleNegativesRankingLoss",
      "matryoshka_dims": [
          768,
          512,
          256,
          128,
          64
      ],
      "matryoshka_weights": [
          1,
          1,
          1,
          1,
          1
      ],
      "n_dims_per_step": -1
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: epoch
• gradient_accumulation_steps: 4
• learning_rate: 1e-05
• weight_decay: 0.01
• num_train_epochs: 4
• lr_scheduler_type: cosine
• warmup_ratio: 0.1
• fp16: True
• load_best_model_at_end: True
• optim: adamw_torch_fused
• push_to_hub: True
• hub_model_id: Shashwat13333/msmarco-distilbert-base-v4_1
• push_to_hub_model_id: msmarco-distilbert-base-v4_1
• batch_sampler: no_duplicates

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: epoch
• prediction_loss_only: True
• per_device_train_batch_size: 8
• per_device_eval_batch_size: 8
• per_gpu_train_batch_size: None
• per_gpu_eval_batch_size: None
• gradient_accumulation_steps: 4
• eval_accumulation_steps: None
• torch_empty_cache_steps: None
• learning_rate: 1e-05
• weight_decay: 0.01
• 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: 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}
• deepspeed: None
• label_smoothing_factor: 0.0
• optim: adamw_torch_fused
• 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: True
• resume_from_checkpoint: None
• hub_model_id: Shashwat13333/msmarco-distilbert-base-v4_1
• 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: msmarco-distilbert-base-v4_1
• 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: no_duplicates
• multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	dim_768_cosine_ndcg@10	dim_512_cosine_ndcg@10	dim_256_cosine_ndcg@10	dim_128_cosine_ndcg@10	dim_64_cosine_ndcg@10
0.2	1	4.0076	-	-	-	-	-
1.0	5	4.8662	0.3288	0.3390	0.3208	0.3246	0.2749
2.0	10	4.1825	0.3288	0.3456	0.3306	0.3405	0.2954
3.0	15	3.048	0.3329	0.3313	0.3346	0.3392	0.3227
4.0	20	2.5029	0.3349	0.3382	0.338	0.3429	0.3229

• The bold row denotes the saved checkpoint.

Framework Versions

• Python: 3.11.11
• Sentence Transformers: 3.3.1
• Transformers: 4.47.1
• PyTorch: 2.5.1+cu124
• Accelerate: 1.2.1
• Datasets: 3.2.0
• Tokenizers: 0.21.0

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",
}

MatryoshkaLoss

@misc{kusupati2024matryoshka,
    title={Matryoshka Representation Learning},
    author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi},
    year={2024},
    eprint={2205.13147},
    archivePrefix={arXiv},
    primaryClass={cs.LG}
}

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}
}