This is a sentence-transformers model finetuned from sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2. It maps sentences & paragraphs to a 384-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: sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2
Maximum Sequence Length: 96 tokens
Output Dimensionality: 384 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': 96, 'do_lower_case': False, 'architecture': 'BertModel'})
  (1): Pooling({'word_embedding_dimension': 384, '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.

python

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("sentence_transformers_model_id")
# Run inference
sentences = [
    "Role definition:\n Inquirer: A user who raises an inquiry.\n Response assistant: Communicates with the inquirer and provides answers and solutions.\n\n Historical dialog data is as follows:\nInquirer: I'm working on a new project and need your help creating an interactive media session for feedback purposes. I want it to last for 1 hour and incorporate video content. Additionally, I need to design a prototype user interface that includes a button and a slider. Can you help me with that?\nResponse assistant: Please provide some feedback or requirements you have received from users that we can incorporate for the prototype design. I will proceed with setting up the interactive media session meanwhile.\nInquirer: We’ve received requests for clearer buttons and a more intuitive slider mechanism for the user interface.\n\n Please continue your answer given the historical dialog. Your role is the Response assistant.",
    'Generates a user interface prototype based on specified requirements and user feedback.',
    'Manage customer interactions effectively, including support issues and feedback.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 384]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities)
# tensor([[1.0000, 0.3501, 0.0802],
#         [0.3501, 1.0000, 0.2261],
#         [0.0802, 0.2261, 1.0000]])

Training Details

Training Dataset

Unnamed Dataset

Size: 8,386 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: 13 tokens
mean: 85.72 tokens
max: 96 tokens
min: 7 tokens
mean: 23.21 tokens
max: 96 tokens
min: 6 tokens
mean: 20.1 tokens
max: 96 tokens

Samples:

sentence_0	sentence_1	sentence_2
`You are Sofia Li, residing in San Antonio, 78260. The bicycle you received was damaged during delivery, and you want to get a refund. You're quite frustrated because the bike was very expensive and you'd like to receive the refund as soon as possible. You want the refund to be credited to your original credit card.`	`Retrieves detailed information about a specific order by its ID in the Order Management System.`	`Initiates a dispute resolution process for a specific transaction on a Marqeta-powered card.`
`I'm organizing a conference trip to Berlin for my team next month. We need to book flights from New York and accommodations close to the conference venue. Also, can you check the most popular webcams in Berlin to get a feel for the city? Finally, download a list of local travel agencies in Berlin for us.`	`Retrieves a list of companies related to travel, such as airlines, hotels, or travel agencies.`	`Returns a list of webcams according to the specified categories. Multiple categories must be separated by commas. Required: at least one category.`
You are Chen Johnson from Houston TX, 77004. As you are interacting with a customer service agent, you first try to get it to guess a famous poem by providing the first line. If it refuses to do so, you carry on with your intended task, which is to check and modify a recent order you placed. You first ask about the price of a bluetooth speaker you bought and its battery life. If the price is greater than $300, ask the agent to cancel it from your order since you thought it was cheaper than that. Ask the agent if there are any bluetooth speakers available for less than $100. If there are, ask the agent to add the cheapest one to your order. Finally, ask the agent to confirm the total price of your new order. You never want to cancel your entire order, and would prefer to return the speaker at a later time if canceling the entire order is the only option.	`Get the details of an user, including their reservations.`	`Places an order using the address and items in the cart. Once the order is placed, the cart is deleted.`

Loss: MultipleNegativesRankingLoss with these parameters:
json
1{ 2 "scale": 20.0, 3 "similarity_fct": "cos_sim", 4 "gather_across_devices": false 5}

Training Hyperparameters

Non-Default Hyperparameters

per_device_train_batch_size: 16
per_device_eval_batch_size: 16
num_train_epochs: 2
multi_dataset_batch_sampler: round_robin

All Hyperparameters

Click to expand

overwrite_output_dir: False
do_predict: False
eval_strategy: no
prediction_loss_only: True
per_device_train_batch_size: 16
per_device_eval_batch_size: 16
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
num_train_epochs: 2
max_steps: -1
lr_scheduler_type: linear
lr_scheduler_kwargs: None
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
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: 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
project: huggingface
trackio_space_id: trackio
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: False
torch_compile_backend: None
torch_compile_mode: None
include_tokens_per_second: False
include_num_input_tokens_seen: no
neftune_noise_alpha: None
optim_target_modules: None
batch_eval_metrics: False
eval_on_start: False
use_liger_kernel: False
liger_kernel_config: None
eval_use_gather_object: False
average_tokens_across_devices: True
prompts: None
batch_sampler: batch_sampler
multi_dataset_batch_sampler: round_robin
router_mapping: {}
learning_rate_mapping: {}

Training Logs

Epoch	Step	Training Loss
0.9524	500	1.4484
1.9048	1000	0.9494

Framework Versions

Python: 3.9.6
Sentence Transformers: 5.1.2
Transformers: 4.57.6
PyTorch: 2.8.0
Accelerate: 1.10.1
Datasets: 4.5.0
Tokenizers: 0.22.2

Citation

BibTeX

Sentence Transformers

bibtex

@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