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.

python

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:
json
1{ 2 "activation_fn": "torch.nn.modules.linear.Identity", 3 "mini_batch_size": null 4}

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:
json
1{ 2 "activation_fn": "torch.nn.modules.linear.Identity", 3 "mini_batch_size": null 4}

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

bibtex

@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

bibtex

@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}
}

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