{{ model_name if model_name else ((model_type or "Sparse Encoder") + " model") }}

This is a {{ model_type or "Sparse Encoder" }} model{% if base_model %} finetuned from [{{ base_model }}](https://huggingface.co/{{ base_model }}){% else %} trained{% endif %}{% if train_datasets | selectattr("name") | list %} on {% if train_datasets | selectattr("name") | map(attribute="name") | join(", ") | length > 200 %}{{ train_datasets | length }}{% else %}the {% for dataset in (train_datasets | selectattr("name")) %}{% if dataset.id %}[{{ dataset.name if dataset.name else dataset.id }}](https://huggingface.co/datasets/{{ dataset.id }}){% else %}{{ dataset.name }}{% endif %}{% if not loop.last %}{% if loop.index == (train_datasets | selectattr("name") | list | length - 1) %} and {% else %}, {% endif %}{% endif %}{% endfor %}{% endif %} dataset{{"s" if train_datasets | selectattr("name") | list | length > 1 else ""}}{% endif %} using the sentence-transformers library. It maps sentences & paragraphs to a {{ output_dimensionality }}-dimensional sparse vector space {% if "CSR" in model_type %} with {{ max_active_dims }} maximum active dimensions{% endif %} and can be used for {{ task_name }}.

Model Details

Model Description

• Model Type: {{ model_type or "Sparse Encoder" }} {% if base_model -%} {%- if base_model_revision -%}
  • Base model: [{{ base_model }}](https://huggingface.co/{{ base_model }}) <!-- at revision {{ base_model_revision }} --> {%- else -%}
  • Base model: [{{ base_model }}](https://huggingface.co/{{ base_model }}) {%- endif -%} {%- else -%}
  <!-- - **Base model:** [Unknown](https://huggingface.co/unknown) -->

{%- endif %}

• Maximum Sequence Length: {{ model_max_length }} tokens
• Output Dimensionality: {{ output_dimensionality }} dimensions{% if max_active_dims %} (trained with {{ max_active_dims }} maximum active dimensions){% endif %}
• Similarity Function: {{ similarity_fn_name }} {% if supported_modalities -%}
  • Supported Modalit{{"ies" if supported_modalities | length > 1 else "y"}}: {{ supported_modalities | join(", ") }} {%- endif %} {% if train_datasets | selectattr("name") | list -%}
  • Training Dataset{{"s" if train_datasets | selectattr("name") | list | length > 1 else ""}}: {%- for dataset in (train_datasets | selectattr("name")) %} {%- if dataset.id %}
  • [{{ dataset.name if dataset.name else dataset.id }}](https://huggingface.co/datasets/{{ dataset.id }}) {%- else %}
  • {{ dataset.name }} {%- endif %} {%- endfor %} {%- else -%}
  <!-- - **Training Dataset:** Unknown -->

{%- endif %} {% if language -%} - Language{{"s" if language is not string and language | length > 1 else ""}}: {%- if language is string %} {{ language }} {%- else %} {% for lang in language -%} {{ lang }}{{ ", " if not loop.last else "" }} {%- endfor %} {%- endif %} {%- else -%} <!-- - **Language:** Unknown --> {%- endif %} {% if license -%} - License: {{ license }} {%- else -%} <!-- - **License:** Unknown --> {%- endif %}

Model Sources

• Documentation: Sentence Transformers Documentation
• Documentation: Sparse Encoder Documentation
• Repository: Sentence Transformers on GitHub
• Hugging Face: Sparse Encoders on Hugging Face

Full Model Architecture

{{ model_string }}

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. {{ usage_snippet }}

{% if eval_metrics %}

Evaluation

Metrics

{% for metrics in eval_metrics %}

{{ metrics.description }}

{% if metrics.dataset_name %}

• Dataset{% if metrics.dataset_name is not string and metrics.dataset_name | length > 1 %}s{% endif %}: {% if metrics.dataset_name is string -%} {{ metrics.dataset_name }} {%- else -%} {%- for name in metrics.dataset_name -%} {{ name }} {%- if not loop.last -%} {%- if loop.index == metrics.dataset_name | length - 1 %} and {% else -%}, {% endif -%} {%- endif -%} {%- endfor -%} {%- endif -%} {%- endif %}
• Evaluated with {% if metrics.class_name.startswith("sentence_transformers.") %}[<code>{{ metrics.class_name.split(".")[-1] }}</code>](https://sbert.net/docs/package_reference/sparse_encoder/evaluation.html#sentence_transformers.sparse_encoder.evaluation.{{ metrics.class_name.split(".")[-1] }}){% else %}<code>{{ metrics.class_name }}</code>{% endif %}{% if metrics.config_code %} with these parameters: {{ metrics.config_code }}{% endif %}

{{ metrics.table }} {%- endfor %}{% endif %}

Training Details

{% for dataset_type, dataset_list in [("training", train_datasets), ("evaluation", eval_datasets)] %}{% if dataset_list %}

{{ dataset_type.title() }} Dataset{{"s" if dataset_list | length > 1 else ""}}

{% for dataset in dataset_list %}{% if dataset_list | length > 3 %}<details><summary>{{ dataset['name'] or 'Unnamed Dataset' }}</summary> {% endif %}

{{ dataset['name'] or 'Unnamed Dataset' }}

{% if dataset['name'] %}

• Dataset: {% if 'id' in dataset %}[{{ dataset['name'] }}](https://huggingface.co/datasets/{{ dataset['id'] }}){% else %}{{ dataset['name'] }}{% endif %} {%- if 'revision' in dataset and 'id' in dataset %} at [{{ dataset['revision'][:7] }}](https://huggingface.co/datasets/{{ dataset['id'] }}/tree/{{ dataset['revision'] }}){% endif %}{% endif %} {% if dataset['size'] %}* Size: {{ "{:,}".format(dataset['size']) }} {{ dataset_type }} samples {% endif %}* Columns: {% if dataset['columns'] | length == 1 %}{{ dataset['columns'][0] }}{% elif dataset['columns'] | length == 2 %}{{ dataset['columns'][0] }} and {{ dataset['columns'][1] }}{% else %}{{ dataset['columns'][:-1] | join(', ') }}, and {{ dataset['columns'][-1] }}{% endif %} {% if dataset['stats_table'] %}* Approximate statistics based on the first {{ [dataset['size'], 1000] | min }} samples: {{ dataset['stats_table'] }}{% endif %}{% if dataset['examples_table'] %}* Samples: {{ dataset['examples_table'] }}{% endif %}* Loss: {% if dataset["loss"]["fullname"].startswith("sentence_transformers.") %}[<code>{{ dataset["loss"]["fullname"].split(".")[-1] }}</code>](https://sbert.net/docs/package_reference/sparse_encoder/losses.html#{{ dataset["loss"]["fullname"].split(".")[-1].lower() }}){% else %}<code>{{ dataset["loss"]["fullname"] }}</code>{% endif %}{% if "config_code" in dataset["loss"] %} with these parameters: {{ dataset["loss"]["config_code"] }}{% endif %} {% if dataset_list | length > 3 %}</details> {% endif %}{% endfor %}{% endif %}{% endfor -%}

{% if all_hyperparameters %}

Training Hyperparameters

{% if non_default_hyperparameters -%}

Non-Default Hyperparameters

{% for name, value in non_default_hyperparameters.items() %}- {{ name }}: {{ value }} {% endfor %}{%- endif %}

All Hyperparameters

{% for name, value in all_hyperparameters.items() %}- {{ name }}: {{ value }} {% endfor %}

{%- if eval_lines %}

Training Logs

{% if hide_eval_lines %}<details><summary>Click to expand</summary>

{% endif -%} {{ eval_lines }}{% if explain_bold_in_eval %}

• The bold row denotes the saved checkpoint.{% endif %} {%- if hide_eval_lines %}

{%- if co2_eq_emissions %}

Environmental Impact

Carbon emissions were measured using CodeCarbon.

• Energy Consumed: {{ "%.3f"|format(co2_eq_emissions["energy_consumed"]) }} kWh
• Carbon Emitted: {{ "%.3f"|format(co2_eq_emissions["emissions"] / 1000) }} kg of CO2

Training Hardware

• On Cloud: {{ "Yes" if co2_eq_emissions["on_cloud"] else "No" }}
• GPU Model: {{ co2_eq_emissions["hardware_used"] or "No GPU used" }}
• CPU Model: {{ co2_eq_emissions["cpu_model"] }}
• RAM Size: {{ "%.2f"|format(co2_eq_emissions["ram_total_size"]) }} GB {% endif %} {%- if training_time is not none %}

Training Time

• Training: {{ training_time }} {%- if evaluation_time is not none %}
• Evaluation: {{ evaluation_time }}
• Total: {{ total_time }} {%- endif %} {% endif %}

Framework Versions

• Python: {{ version["python"] }}
• Sentence Transformers: {{ version["sentence_transformers"] }}
• Transformers: {{ version["transformers"] }}
• PyTorch: {{ version["torch"] }}
• Accelerate: {{ version["accelerate"] }}
• Datasets: {{ version["datasets"] }}
• Tokenizers: {{ version["tokenizers"] }}

Citation

BibTeX

{% for loss_name, citation in citations.items() %}

{{ loss_name }}

{{ citation | trim }}

{% endfor %}