Fine-tuned ColBERT model for semantic caching

This is a PyLate model finetuned from lightonai/GTE-ModernColBERT-v1 on the LangCache Sentence Pairs (subsets=['all'], train+val=True) dataset. It maps sentences & paragraphs to sequences of 128-dimensional dense vectors and can be used for semantic textual similarity using the MaxSim operator.

Model Details

Model Description

Model Type: PyLate model
Base model: lightonai/GTE-ModernColBERT-v1
Document Length: 512 tokens
Query Length: 512 tokens
Output Dimensionality: 128 tokens
Similarity Function: MaxSim
Training Dataset:
- LangCache Sentence Pairs (subsets=['all'], train+val=True)
Language: en
License: apache-2.0

Model Sources

Documentation: PyLate Documentation
Repository: PyLate on GitHub
Hugging Face: PyLate models on Hugging Face

Full Model Architecture

ColBERT(
  (0): Transformer({'max_seq_length': 511, 'do_lower_case': False, 'architecture': 'ModernBertModel'})
  (1): Dense({'in_features': 768, 'out_features': 128, 'bias': False, 'activation_function': 'torch.nn.modules.linear.Identity', 'use_residual': False})
)

Usage

First install the PyLate library:

pip install -U pylate

Retrieval

Use this model with PyLate to index and retrieve documents. The index uses FastPLAID for efficient similarity search.

Indexing documents

Load the ColBERT model and initialize the PLAID index, then encode and index your documents:

from pylate import indexes, models, retrieve

# Step 1: Load the ColBERT model
model = models.ColBERT(
    model_name_or_path="aditeyabaral/semcache-colbert-v1",
)

# Step 2: Initialize the PLAID index
index = indexes.PLAID(
    index_folder="pylate-index",
    index_name="index",
    override=True,  # This overwrites the existing index if any
)

# Step 3: Encode the documents
documents_ids = ["1", "2", "3"]
documents = ["document 1 text", "document 2 text", "document 3 text"]

documents_embeddings = model.encode(
    documents,
    batch_size=32,
    is_query=False,  # Ensure that it is set to False to indicate that these are documents, not queries
    show_progress_bar=True,
)

# Step 4: Add document embeddings to the index by providing embeddings and corresponding ids
index.add_documents(
    documents_ids=documents_ids,
    documents_embeddings=documents_embeddings,
)

Note that you do not have to recreate the index and encode the documents every time. Once you have created an index and added the documents, you can re-use the index later by loading it:

# To load an index, simply instantiate it with the correct folder/name and without overriding it
index = indexes.PLAID(
    index_folder="pylate-index",
    index_name="index",
)

Retrieving top-k documents for queries

Once the documents are indexed, you can retrieve the top-k most relevant documents for a given set of queries. To do so, initialize the ColBERT retriever with the index you want to search in, encode the queries and then retrieve the top-k documents to get the top matches ids and relevance scores:

# Step 1: Initialize the ColBERT retriever
retriever = retrieve.ColBERT(index=index)

# Step 2: Encode the queries
queries_embeddings = model.encode(
    ["query for document 3", "query for document 1"],
    batch_size=32,
    is_query=True,  #  # Ensure that it is set to False to indicate that these are queries
    show_progress_bar=True,
)

# Step 3: Retrieve top-k documents
scores = retriever.retrieve(
    queries_embeddings=queries_embeddings,
    k=10,  # Retrieve the top 10 matches for each query
)

Reranking

If you only want to use the ColBERT model to perform reranking on top of your first-stage retrieval pipeline without building an index, you can simply use rank function and pass the queries and documents to rerank:

from pylate import rank, models

queries = [
    "query A",
    "query B",
]

documents = [
    ["document A", "document B"],
    ["document 1", "document C", "document B"],
]

documents_ids = [
    [1, 2],
    [1, 3, 2],
]

model = models.ColBERT(
    model_name_or_path="aditeyabaral/semcache-colbert-v1",
)

queries_embeddings = model.encode(
    queries,
    is_query=True,
)

documents_embeddings = model.encode(
    documents,
    is_query=False,
)

reranked_documents = rank.rerank(
    documents_ids=documents_ids,
    queries_embeddings=queries_embeddings,
    documents_embeddings=documents_embeddings,
)

Evaluation

Metrics

Col BERTTriplet

Dataset: test_triplet
Evaluated with pylate.evaluation.colbert_triplet.ColBERTTripletEvaluator

Metric	Value
accuracy	0.9079

Training Details

Training Dataset

LangCache Sentence Pairs (subsets=['all'], train+val=True)

Dataset: LangCache Sentence Pairs (subsets=['all'], train+val=True)
Size: 1,452,533 training samples
Columns: anchor, positive, and negative_1

Approximate statistics based on the first 1000 samples:

	anchor	positive	negative_1
type	string	string	string
details	min: 9 tokens mean: 28.67 tokens max: 79 tokens	min: 8 tokens mean: 28.51 tokens max: 57 tokens	min: 5 tokens mean: 24.02 tokens max: 50 tokens

Samples:

anchor	positive	negative_1
`Any Canadian teachers (B.Ed. holders) teaching in U.S. schools?`	`Any Canadian teachers (B.Ed. holders) teaching in U.S. schools?`	`Are there many Canadians living and working illegally in the United States?`
`Are there any underlying psychological tricks/tactics that are used when designing the lines for rides at amusement parks?`	`Are there any underlying psychological tricks/tactics that are used when designing the lines for rides at amusement parks?`	`Is there any tricks for straight lines mcqs?`
`Can I pay with a debit card on PayPal?`	`Can I pay with a debit card on PayPal?`	`Can you transfer PayPal funds onto a debit card/credit card?`

Loss: pylate.losses.contrastive.Contrastive

Evaluation Dataset

LangCache Sentence Pairs (split=test)

Dataset: LangCache Sentence Pairs (split=test)
Size: 110,066 evaluation samples
Columns: anchor, positive, and negative_1

Approximate statistics based on the first 1000 samples:

	anchor	positive	negative_1
type	string	string	string
details	min: 5 tokens mean: 26.68 tokens max: 104 tokens	min: 5 tokens mean: 26.34 tokens max: 104 tokens	min: 6 tokens mean: 20.39 tokens max: 69 tokens

Samples:

anchor	positive	negative_1
`What high potential jobs are there other than computer science?`	`What high potential jobs are there other than computer science?`	`Why IT or Computer Science jobs are being over rated than other Engineering jobs?`
`Would India ever be able to develop a missile system like S300 or S400 missile?`	`Would India ever be able to develop a missile system like S300 or S400 missile?`	`Should India buy the Russian S400 air defence missile system?`
`water from the faucet is being drunk by a yellow dog`	`A yellow dog is drinking water from the faucet`	`Do you get more homework in 9th grade than 8th?`

Loss: pylate.losses.contrastive.Contrastive

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: steps
per_device_train_batch_size: 48
per_device_eval_batch_size: 48
learning_rate: 0.0002
weight_decay: 0.001
num_train_epochs: 5
warmup_ratio: 0.1
load_best_model_at_end: True
optim: adamw_torch
ddp_find_unused_parameters: False
push_to_hub: True
hub_model_id: aditeyabaral/semcache-colbert-v1
eval_on_start: True
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: 48
per_device_eval_batch_size: 48
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: 0.0002
weight_decay: 0.001
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: 0.1
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: 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: True
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
optim_args: None
adafactor: False
group_by_length: False
length_column_name: length
ddp_find_unused_parameters: False
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: aditeyabaral/semcache-colbert-v1
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: True
use_liger_kernel: False
liger_kernel_config: None
eval_use_gather_object: False
average_tokens_across_devices: True
prompts: None
batch_sampler: no_duplicates
multi_dataset_batch_sampler: proportional
router_mapping: {}
learning_rate_mapping: {}

Training Logs

Epoch	Step	Training Loss	Validation Loss	accuracy
0	0	-	1.5292	0.9079
0.1322	1000	1.3853	-	-
0.2644	2000	0.5592	-	-
0.3966	3000	0.409	-	-
0.5288	4000	0.3585	-	-
0.6609	5000	0.3153	-	-
0.7931	6000	0.2972	-	-
0.9253	7000	0.2738	-	-
1.0575	8000	0.2544	-	-
1.1897	9000	0.2476	-	-
1.3219	10000	0.2188	-	-
1.4541	11000	0.1609	-	-
1.5863	12000	0.1299	-	-
1.7184	13000	0.1343	-	-
1.8506	14000	0.1288	-	-
1.9828	15000	0.1371	-	-
2.1150	16000	0.1187	-	-
2.2472	17000	0.1237	-	-
2.3794	18000	0.1152	-	-
2.5116	19000	0.0837	-	-
2.6438	20000	0.086	-	-
2.7759	21000	0.0821	-	-
2.9081	22000	0.0894	-	-
3.0403	23000	0.0857	-	-
3.1725	24000	0.0898	-	-
3.3047	25000	0.0817	-	-
3.4369	26000	0.0714	-	-
3.5691	27000	0.0622	-	-
3.7013	28000	0.0682	-	-
3.8334	29000	0.0704	-	-
3.9656	30000	0.0738	-	-
4.0978	31000	0.0693	-	-
4.2300	32000	0.0808	-	-
4.3622	33000	0.0726	-	-
4.4944	34000	0.0584	-	-
4.6266	35000	0.0557	-	-
4.7588	36000	0.0615	-	-
4.8909	37000	0.0656	-	-

Framework Versions

Python: 3.12.12
Sentence Transformers: 5.1.1
PyLate: 1.4.0
Transformers: 4.56.2
PyTorch: 2.9.0+cu128
Accelerate: 1.13.0
Datasets: 4.6.1
Tokenizers: 0.22.2

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"
}

PyLate

@inproceedings{DBLP:conf/cikm/ChaffinS25,
  author       = {Antoine Chaffin and
                  Rapha{"{e}}l Sourty},
  editor       = {Meeyoung Cha and
                  Chanyoung Park and
                  Noseong Park and
                  Carl Yang and
                  Senjuti Basu Roy and
                  Jessie Li and
                  Jaap Kamps and
                  Kijung Shin and
                  Bryan Hooi and
                  Lifang He},
  title        = {PyLate: Flexible Training and Retrieval for Late Interaction Models},
  booktitle    = {Proceedings of the 34th {ACM} International Conference on Information
                  and Knowledge Management, {CIKM} 2025, Seoul, Republic of Korea, November
                  10-14, 2025},
  pages        = {6334--6339},
  publisher    = {{ACM}},
  year         = {2025},
  url          = {https://github.com/lightonai/pylate},
  doi          = {10.1145/3746252.3761608},
}