bge_small_xtr_contrastive_k64

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 <!-- at revision 5c38ec7c405ec4b44b94cc5a9bb96e735b38267a -->
• Document Length: 300 tokens
• Query Length: 32 tokens
• Output Dimensionality: 128 tokens
• Similarity Function: MaxSim
• Training Dataset:
  • msmarco-10m-triplets

<!-- - **Language:** Unknown --> <!-- - **License:** Unknown -->

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,
)

<!-- ### Direct Usage (Transformers) <details><summary>Click to see the direct usage in Transformers</summary> </details> --> <!-- ### Downstream Usage (Sentence Transformers) You can finetune this model on your own dataset. <details><summary>Click to expand</summary> </details> --> <!-- ### Out-of-Scope Use *List how the model may foreseeably be misused and address what users ought not to do with the model.* -->

Evaluation

Metrics

Col BERTTriplet

• Evaluated with <code>pylate.evaluation.colbert_triplet.ColBERTTripletEvaluator</code>

<!-- ## Bias, Risks and Limitations *What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.* --> <!-- ### Recommendations *What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.* -->

Training Details

Training Dataset

msmarco-10m-triplets

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

  	query	positive	negative
  type	string	string	string
  details	<ul><li>min: 32 tokens</li><li>mean: 32.0 tokens</li><li>max: 32 tokens</li></ul>	<ul><li>min: 32 tokens</li><li>mean: 32.0 tokens</li><li>max: 32 tokens</li></ul>	<ul><li>min: 32 tokens</li><li>mean: 32.0 tokens</li><li>max: 32 tokens</li></ul>

• Samples:

  query	positive	negative
  <code>what causes infertility in males</code>	<code>Infertility in men. Semen and sperm. The most common cause of infertility in men is poor quality semen, the fluid containing sperm that's ejaculated during sex. Possible reasons for abnormal semen include: a lack of sperm – you may have a very low sperm count, or no sperm at all.</code>	<code>Only two species have been observed showing a same-sex preference for life, even when partners of the opposite sex are available. One is, of course, humans. The other is domestic sheep. In flocks of sheep, up to 8% of the males prefer other males even when fertile females are around.</code>
  <code>what is an authentication method</code>	<code>Authentication is a process in which the credentials provided are compared to those on file in a database of authorized users’ information on a local operating system or within an authentication server. If the credentials match, the process is completed and the user is granted authorization for access.</code>	<code>Using the Graphical User Interface. Cisco Prime Access Registrar (Prime Access Registrar) is a Remote Authentication Dial-In User Service. (RADIUS) server that enables multiple dial-in Ne twork Access Server (NAS) devices to share a common. authentication, authorization, and accounting database. This chapter describes how to use the standalone graphical user interface (GUI) of.</code>
  <code>male seminal vesicles what they do and what they are</code>	<code>Seminal gland. The seminal gland, more commonly referred to as the seminal vesicle, holds the liquid that mixes with sperm to form semen. Semen combines fluid elements from the epididymis, seminal vesicles, prostate gland, and vas deferens.Each body part plays a key role in semen production.The fluids help the sperm swim towards the egg and keep the sperm nourished during the transit process.eminal gland. The seminal gland, more commonly referred to as the seminal vesicle, holds the liquid that mixes with sperm to form semen. Semen combines fluid elements from the epididymis, seminal vesicles, prostate gland, and vas deferens.</code>	<code>Key points. 1 Vesicles derived from the outer membrane of Gram-negative bacteria, or outer-membrane vesicles (OMVs), are heterogeneous in size and composition, encapsulate soluble periplasmic content and are ubiquitously produced.</code>

• Loss: <code>pylate.losses.xtr_primeqa.XTRPrimeQA</code>

Evaluation Dataset

msmarco-10m-triplets

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

  	query	positive	negative
  type	string	string	string
  details	<ul><li>min: 32 tokens</li><li>mean: 32.0 tokens</li><li>max: 32 tokens</li></ul>	<ul><li>min: 32 tokens</li><li>mean: 32.0 tokens</li><li>max: 32 tokens</li></ul>	<ul><li>min: 32 tokens</li><li>mean: 32.0 tokens</li><li>max: 32 tokens</li></ul>

• Samples:

  query	positive	negative
  <code>definition of muse</code>	<code>Muse(noun) a gap or hole in a hedge, hence, wall, or the like, through which a wild animal is accustomed to pass; a muset. Muse(noun) one of the nine goddesses who presided over song and the different kinds of poetry, and also the arts and sciences; -- often used in the plural. Muse(noun) a particular power and practice of poetry. Muse(noun) a poet; a bard. Muse(noun) to think closely; to study in silence; to meditate.</code>	<code>by definition. : because of what something or someone is : according to the definition of a word that is being used to describe someone or something. A volunteer by definition is not paid. A glider is by definition an aircraft with no engine.</code>
  <code>are marines first to be deployed in wartime</code>	<code>Best Answer: Well the Marines deploy regardless of war or no war.. Before a war happens the Marines are already in the area..They are always deployed out a sea awaiting the next crisis..</code>	<code>How the U.S. Marine Corps Was Founded Twice. On Tuesday, the U.S. Marine Corps celebrates its 240th birthday, marking the Nov. 10, 1775, decision by the Second Continental Congress to establish two battalions of Marines. Except those weren't the Marines.</code>
  <code>are hybrid long term care policy premiums tax deductible</code>	<code>1 Premiums paid on a long-term care insurance product may be eligible for an income tax deduction. 2 The amount of the deduction depends on the age of the covered person. 3 Benefits paid from a long-term care contract are generally excluded from income. 4 Business deductions of premiums are determined by the type of business.</code>	<code>An HMO plan may be right for you if: 1 You're shopping for a plan with lower premiums. 2 You want a plan without a deductible and don't mind having an out-of-pocket limit. 3 You need preventive care services such as coverage for checkups and immunizations.</code>

• Loss: <code>pylate.losses.xtr_primeqa.XTRPrimeQA</code>

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

<details><summary>Click to expand</summary>

• 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: {}

</details>