Pranjal2002/finetuned_colbert_finance_v2

CrossEncoder based on colbert-ir/colbertv2.0

This is a Cross Encoder model finetuned from colbert-ir/colbertv2.0 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: colbert-ir/colbertv2.0
• Maximum Sequence Length: 512 tokens
• Number of Output Labels: 1 label

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("Pranjal2002/finetuned_colbert_finance_v2")
# Get scores for pairs of texts
pairs = [
    ['What guidance was offered on The Estée Lauder Companies Inc.’s inventory management or supply chain efficiency targets?', 'Earnings'],
    ['What guidance was offered on The Estée Lauder Companies Inc.’s inventory management or supply chain efficiency targets?', '8-K'],
    ['What guidance was offered on The Estée Lauder Companies Inc.’s inventory management or supply chain efficiency targets?', 'DEF14A'],
    ['What guidance was offered on The Estée Lauder Companies Inc.’s inventory management or supply chain efficiency targets?', '10-K'],
    ['What guidance was offered on The Estée Lauder Companies Inc.’s inventory management or supply chain efficiency targets?', '10-Q'],
]
scores = model.predict(pairs)
print(scores.shape)
# (5,)

# Or rank different texts based on similarity to a single text
ranks = model.rank(
    'What guidance was offered on The Estée Lauder Companies Inc.’s inventory management or supply chain efficiency targets?',
    [
        'Earnings',
        '8-K',
        'DEF14A',
        '10-K',
        '10-Q',
    ]
)
# [{'corpus_id': ..., 'score': ...}, {'corpus_id': ..., 'score': ...}, ...]

Training Details

Training Dataset

Unnamed Dataset

• Size: 3,988 training samples
• Columns: query, docs, and labels
• Approximate statistics based on the first 1000 samples:

  	query	docs	labels
  type	string	list	list
  details	min: 53 characters mean: 101.87 characters max: 197 characters	size: 5 elements	size: 5 elements

• Samples:

  query	docs	labels
  How has Keurig Dr Pepper’s beverage segment profitability trended over recent periods?	['10-Q', '10-K', 'Earnings', '8-K', 'DEF14A']	[4, 3, 2, 1, 0]
  How does management describe competitive advantages in generative AI developer tooling	['Earnings', '10-K', 'DEF14A', '8-K', '10-Q']	[4, 3, 2, 1, 0]
  What did Mohawk Industries’ leadership say about Mohawk Industries’ share repurchase plans?	['10-K', '10-Q', 'Earnings', 'DEF14A', '8-K']	[2, 2, 1, 0, 0]

• Loss: ListNetLoss with these parameters:

  {
      "activation_fn": "torch.nn.modules.linear.Identity",
      "mini_batch_size": null
  }
  

Evaluation Dataset

Unnamed Dataset

• Size: 998 evaluation samples
• Columns: query, docs, and labels
• Approximate statistics based on the first 998 samples:

  	query	docs	labels
  type	string	list	list
  details	min: 43 characters mean: 102.97 characters max: 203 characters	size: 5 elements	size: 5 elements

• Samples:

  query	docs	labels
  What guidance was offered on The Estée Lauder Companies Inc.’s inventory management or supply chain efficiency targets?	['Earnings', '8-K', 'DEF14A', '10-K', '10-Q']	[4, 3, 2, 1, 0]
  What questions were asked about Live Nation Entertainment’s concert attendance and ticket sales engagement metrics?	['Earnings', '10-K', '8-K', '10-Q', 'DEF14A']	[4, 3, 2, 1, 0]
  How has the ratio of AvalonBay Communities’ recurring to one-time rental income evolved in the latest reporting period?	['10-Q', '10-K', 'Earnings', '8-K', 'DEF14A']	[4, 3, 2, 1, 0]

• Loss: ListNetLoss with these parameters:

  {
      "activation_fn": "torch.nn.modules.linear.Identity",
      "mini_batch_size": null
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 4
• per_device_eval_batch_size: 4
• gradient_accumulation_steps: 2
• learning_rate: 2e-05
• num_train_epochs: 5
• warmup_steps: 100
• fp16: True
• load_best_model_at_end: 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: 4
• per_device_eval_batch_size: 4
• per_gpu_train_batch_size: None
• per_gpu_eval_batch_size: None
• gradient_accumulation_steps: 2
• 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: 5
• max_steps: -1
• lr_scheduler_type: linear
• lr_scheduler_kwargs: {}
• warmup_ratio: 0.0
• warmup_steps: 100
• 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}
• parallelism_config: 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: False
• resume_from_checkpoint: None
• hub_model_id: None
• 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: False
• neftune_noise_alpha: None
• optim_target_modules: None
• batch_eval_metrics: False
• eval_on_start: False
• use_liger_kernel: False
• liger_kernel_config: None
• 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

Epoch	Step	Training Loss	Validation Loss
0.1003	50	1.5717	-
0.2006	100	1.4575	-
0.3009	150	1.4404	-
0.4012	200	1.408	1.3705
0.5015	250	1.3936	-
0.6018	300	1.3719	-
0.7021	350	1.3777	-
0.8024	400	1.3689	1.3444
0.9027	450	1.3612	-
1.0020	500	1.3263	-
1.1023	550	1.3493	-
1.2026	600	1.3602	1.3374
1.3029	650	1.3181	-
1.4032	700	1.3217	-
1.5035	750	1.3431	-
1.6038	800	1.3234	1.3374
1.7041	850	1.3317	-
1.8044	900	1.34	-
1.9047	950	1.3467	-
2.0040	1000	1.3236	1.3325
2.1043	1050	1.2743	-
2.2046	1100	1.3177	-
2.3049	1150	1.3004	-
2.4052	1200	1.3114	1.3274
2.5055	1250	1.3138	-
2.6058	1300	1.3263	-
2.7061	1350	1.3175	-
2.8064	1400	1.3033	1.3462
2.9067	1450	1.3112	-
3.0060	1500	1.3025	-
3.1063	1550	1.2818	-
3.2066	1600	1.2768	1.3426
3.3069	1650	1.275	-
3.4072	1700	1.3024	-
3.5075	1750	1.2765	-
3.6078	1800	1.2932	1.3467
3.7081	1850	1.2774	-
3.8084	1900	1.2759	-
3.9087	1950	1.2991	-
4.0080	2000	1.2763	1.3368
4.1083	2050	1.253	-
4.2086	2100	1.243	-
4.3089	2150	1.2719	-
4.4092	2200	1.256	1.3448
4.5095	2250	1.2718	-
4.6098	2300	1.2536	-
4.7101	2350	1.2696	-
4.8104	2400	1.2626	1.3456
4.9107	2450	1.2736	-

• The bold row denotes the saved checkpoint.

Framework Versions

• Python: 3.12.11
• Sentence Transformers: 5.1.0
• Transformers: 4.56.1
• PyTorch: 2.8.0+cu126
• Accelerate: 1.10.1
• Datasets: 4.0.0
• Tokenizers: 0.22.0

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",
}

ListNetLoss

@inproceedings{cao2007learning,
    title={Learning to Rank: From Pairwise Approach to Listwise Approach},
    author={Cao, Zhe and Qin, Tao and Liu, Tie-Yan and Tsai, Ming-Feng and Li, Hang},
    booktitle={Proceedings of the 24th international conference on Machine learning},
    pages={129--136},
    year={2007}
}