xtr-replicability/bge_small_xtr_contrastive_k128

PyLate model based on BAAI/bge-small-en-v1.5

This is a PyLate model finetuned from BAAI/bge-small-en-v1.5 on the msmarco-10m-triplets 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: BAAI/bge-small-en-v1.5
• Document Length: 300 tokens
• Query Length: 32 tokens
• Output Dimensionality: 128 tokens
• Similarity Function: MaxSim
• Training Dataset:
  • msmarco-10m-triplets

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': 300, 'do_lower_case': True, 'architecture': 'BertModel'})
  (1): Dense({'in_features': 384, '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="pylate_model_id",
)

# 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="pylate_model_id",
)

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

• Evaluated with pylate.evaluation.colbert_triplet.ColBERTTripletEvaluator

Metric	Value
accuracy	0.992

Training Details

Training Dataset

msmarco-10m-triplets

• Dataset: msmarco-10m-triplets at 8c5139a
• Size: 9,998,000 training samples
• Columns: query, positive, and negative
• Approximate statistics based on the first 1000 samples:

  	query	positive	negative
  type	string	string	string
  details	min: 32 tokens mean: 32.0 tokens max: 32 tokens	min: 32 tokens mean: 32.0 tokens max: 32 tokens	min: 32 tokens mean: 32.0 tokens max: 32 tokens

• Samples:

  query	positive	negative
  uploading definition	Uploading is the process of transferring data from a computer to a central server; downloading is the opposite. When uploading...	Here are pictures of people with the name Priya. Help us put a face to the name by uploading your pictures to BabyNames.com!ere are pictures of people with the name Priya. Help us put a face to the name by uploading your pictures to BabyNames.com!
  what does it mean theoretical framework	A theoretical framework is a collection of interrelated concepts, like a theory but not necessarily so well worked-out. A theoretical framework guides your research, determining what things you will measure, and what statistical relationships you will look for.	Rails is a software framework for creating web applications. Many students new to programming aren’t sure what the difference is between Ruby and Ruby on Rails. Ruby is a language. Ruby on Rails is a framework that uses and depends on the Ruby programming language. Think of a framework as a collection of pre-written code to help programmers. In this case, the Rails framework is a collection of pre-written code to help programmers make websites and web applications.
  how long is rotisserie chicken good in refrigerator	1 Rotisserie chicken: 3 to 4 days (refrigerator) or 2 to 3 months (freezer). 2 Opened package of hot dogs: 1 week (refrigerator) or 1 to 2 months (freezer). 3 Opened package of deli meat: 3 to 4 days (refrigerator) or 1 to 2 months (freezer).	You can keep cooked chicken in the refrigerator for up to four days. Wrap the cooked chicken in plastic wrap or place it in an airtight container. Put it in the refrigerator as quickly as possible after it cools to room temperature. Ideally, you should store it at the back of the middle shelf in your refrigerator. You can also store multi-ingredient dishes containing cooked chicken, such as chicken salad or chicken casserole, for up to four days.

• Loss: pylate.losses.xtr_primeqa.XTRPrimeQA

Evaluation Dataset

msmarco-10m-triplets

• Dataset: msmarco-10m-triplets at 8c5139a
• Size: 2,000 evaluation samples
• Columns: query, positive, and negative
• Approximate statistics based on the first 1000 samples:

  	query	positive	negative
  type	string	string	string
  details	min: 32 tokens mean: 32.0 tokens max: 32 tokens	min: 32 tokens mean: 32.0 tokens max: 32 tokens	min: 32 tokens mean: 32.0 tokens max: 32 tokens

• Samples:

  query	positive	negative
  what is causing the rise in autoimmune disorders in the us	Furthermore, like chloride and lithium, fluoride is able to displace iodine, contributing to hypothyroidism. Mercury, Nickel and Other Metals. Mercury and nickel have been found to trigger autoimmune thyroid disorders and other autoimmune disorders such as systemic lupus.	Treatments for primary immunodeficiency involve preventing and treating infections, boosting the immune system, and treating the underlying cause of the immune problem. In some cases, primary immune disorders are linked to a serious illness, such as an autoimmune disorder or cancer, which also needs to be treated.
  education expenses tax deductible for real estate agents	Typically, real estate agents may deduct advertising costs, professional and licensing fees, educational costs, a portion of the expenses associated with the business use of their homes and any automobile expenses associated with business use.	Nonbusiness deductions can still result in an NOL: those can include losses due to moving expenses, rental real estate expenses, or casualty and theft losses. But here’s what The Times was getting at: under existing tax laws, if you have an NOL, you first carry back the entire NOL amount to the two prior tax years.
  where is oberlin university	Not to be confused with Oberlin College. J. F. Oberlin University (桜美林大学, Ōbirin daigaku) is a private university in Machida, Tokyo, Japan. The university was founded by Yasuzo Shimizu.	1 Albert Einstein teaching a class of Lincoln University students in University Hall. 3 May, 1946: 6. Albert Einstein with children of Lincoln University faculty at a reception in University Hall.

• Loss: pylate.losses.xtr_primeqa.XTRPrimeQA

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 196
• per_device_eval_batch_size: 196
• learning_rate: 3e-05
• max_grad_norm: 10.0
• num_train_epochs: 0
• max_steps: 50000
• warmup_ratio: 0.01
• bf16: True
• torch_compile: True
• torch_compile_backend: inductor
• eval_on_start: 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: 196
• per_device_eval_batch_size: 196
• 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: 3e-05
• weight_decay: 0.0
• adam_beta1: 0.9
• adam_beta2: 0.999
• adam_epsilon: 1e-08
• max_grad_norm: 10.0
• num_train_epochs: 0
• max_steps: 50000
• lr_scheduler_type: linear
• lr_scheduler_kwargs: {}
• warmup_ratio: 0.01
• 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: True
• 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: 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}
• 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: True
• torch_compile_backend: inductor
• 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: False
• prompts: None
• batch_sampler: batch_sampler
• multi_dataset_batch_sampler: proportional
• router_mapping: {}
• learning_rate_mapping: {}