LorMolf/mnrl-bfcl-roberta-base

SentenceTransformer based on FacebookAI/roberta-base

This is a sentence-transformers model finetuned from FacebookAI/roberta-base. It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.

Model Details

Model Description

• Model Type: Sentence Transformer
• Base model: FacebookAI/roberta-base
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 768 dimensions
• Similarity Function: Cosine Similarity

Model Sources

• Documentation: Sentence Transformers Documentation
• Repository: Sentence Transformers on GitHub
• Hugging Face: Sentence Transformers on Hugging Face

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: RobertaModel 
  (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
)

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.

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("LorMolf/mnrl-bfcl-roberta-base")
# Run inference
sentences = [
    "In a texas holdem game, Who won in the poker game with players Alex, Sam, Robert and Steve given the cards Alex':['A of spades', 'K of spades'], 'Sam': ['2 of diamonds', '3 of clubs'], 'Robert': ['Q of hearts', '10 of hearts'], 'Steve': ['4 of spades', '5 of spades']?",
    'def poker_game_winner:\n\t"""\n\tDescription:\n\t\n\tIdentify the winner in a poker game based on the cards.\n\t\n\tArguments:\n\t---------\n\t- players : array = None (required) Names of the players in a list.\n\t- cards : dict = None (required) An object containing the player name as key and the cards as values in a list.\n\t- type : string = None (optional) Type of poker game. Defaults to \'Texas Holdem\'\n\t\n\t\n\t"""',
    'def geocode_address:\n\t"""\n\tDescription:\n\t\n\tTransforms a description of a location (like a pair of coordinates, an address, or a name of a place) to a location on the Earth\'s surface.\n\t\n\tArguments:\n\t---------\n\t- address : string = None (required) The address that needs to be geocoded.\n\t- locale : string = None (optional) Preferred locale for the returned address information. (Optional) Default: None\n\t\n\t\n\t"""',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]

Evaluation

Metrics

Device Aware Information Retrieval

• Dataset: dev
• Evaluated with src.port.retrieval_evaluator.DeviceAwareInformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.0038
cosine_accuracy@3	0.0095
cosine_accuracy@5	0.0095
cosine_accuracy@10	0.0189
cosine_precision@1	0.0038
cosine_precision@3	0.0032
cosine_precision@5	0.0019
cosine_precision@10	0.0019
cosine_recall@1	0.0038
cosine_recall@3	0.0095
cosine_recall@5	0.0095
cosine_recall@10	0.0189
cosine_ndcg@1	0.0038
cosine_ndcg@3	0.0069
cosine_ndcg@5	0.0069
cosine_ndcg@10	0.01
cosine_mrr@10	0.0074
cosine_map@100	0.0121

Training Details

Training Dataset

Unnamed Dataset

• Size: 3,780 training samples
• Columns: sentence_0, sentence_1, and sentence_2
• Approximate statistics based on the first 1000 samples:

  	sentence_0	sentence_1	sentence_2
  type	string	string	string
  details	min: 5 tokens mean: 46.51 tokens max: 366 tokens	min: 62 tokens mean: 130.97 tokens max: 512 tokens	min: 61 tokens mean: 119.6 tokens max: 428 tokens

• Samples:

  sentence_0	sentence_1	sentence_2
  I want to find 5 restaurants nearby my location, Manhattan, offering Thai food and a vegan menu.	def find_instrument: """ Description: Search for a musical instrument within specified budget and of specific type. Arguments: --------- - budget : float = None (required) Your budget for the instrument. - type : string = None (required) Type of the instrument - make : string = None (optional) Maker of the instrument, Optional parameter. Default is 'all' """	def sports_ranking.get_team_position: """ Description: Retrieve a team's position and stats in the basketball league for a given season. Arguments: --------- - team : string = None (required) The name of the team. - season : string = None (required) The season for which data should be fetched. - detailed : boolean = False (optional) Flag to retrieve detailed stats or just the position. """
  Find the most followed person on twitter who tweets about psychology related to behaviour and group dynamics.	def social_media_analytics.most_followed: """ Description: Find the most followed Twitter user related to certain topics. Arguments: --------- - topic : string = None (required) The main topic of interest. - sub_topics : array = None (optional) Sub-topics related to main topic, Optional. Default is an empty list. - region : string = None (optional) Region of interest for twitter search, Optional. Default is 'global'. """	def library.search_books: """ Description: Search for a book in a given library with optional parameters Arguments: --------- - location : string = None (required) Name or city of library - genre : string = None (required) Genre of the book - title : string = None (optional) Title of the book. Default '' """
  What is the evolutionary history of pandas?	def calculate_biodiversity_index: """ Description: Calculate the biodiversity index of a specific environment or biome using species richness and species evenness. Arguments: --------- - species_richness : integer = None (required) The number of different species in a specific environment. - species_evenness : integer = None (required) The relative abundance of the different species in an environment. - region : string = Desert (optional) The specific environment or biome to be measured. """	def sports_ranking: """ Description: Fetch the ranking of a specific sports team in a specific league Arguments: --------- - team : string = None (required) The name of the team. - league : string = None (required) The name of the league. - season : integer = None (optional) Optional parameter to specify the season, default is the current season, 2024 """

• Loss: MultipleNegativesRankingLoss with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim"
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 2
• per_device_eval_batch_size: 2
• num_train_epochs: 1
• fp16: True
• multi_dataset_batch_sampler: round_robin

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: steps
• prediction_loss_only: True
• per_device_train_batch_size: 2
• per_device_eval_batch_size: 2
• 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: 5e-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: 1
• max_steps: -1
• lr_scheduler_type: linear
• lr_scheduler_kwargs: {}
• warmup_ratio: 0.0
• 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: 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: 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}
• tp_size: 0
• 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}
• 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: 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
• 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
• eval_use_gather_object: False
• average_tokens_across_devices: False
• prompts: None
• batch_sampler: batch_sampler
• multi_dataset_batch_sampler: round_robin

Training Logs

Epoch	Step	Training Loss	dev_cosine_ndcg@10
-1	-1	-	0.1710
0.2	378	-	0.0064
0.2646	500	0.9661	-
0.4	756	-	0.0106
0.5291	1000	1.3862	-
0.6	1134	-	0.0125
0.7937	1500	1.3863	-
0.8	1512	-	0.0106
1.0	1890	-	0.0100

Framework Versions

• Python: 3.10.12
• Sentence Transformers: 4.0.2
• Transformers: 4.51.2
• PyTorch: 2.6.0+cu124
• Accelerate: 1.6.0
• Datasets: 3.5.0
• Tokenizers: 0.21.1

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",
}

MultipleNegativesRankingLoss

@misc{henderson2017efficient,
    title={Efficient Natural Language Response Suggestion for Smart Reply},
    author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil},
    year={2017},
    eprint={1705.00652},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}