This is a sentence-transformers model finetuned from sentence-transformers/all-MiniLM-L6-v2 on the imdb-triplet dataset. 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/all-MiniLM-L6-v2
Maximum Sequence Length: 256 tokens
Output Dimensionality: 384 tokens
Similarity Function: Cosine Similarity
Training Dataset:
- imdb-triplet

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': 256, 'do_lower_case': False}) with Transformer model: 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})
  (2): Normalize()
)

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("celik-muhammed/all-MiniLM-L6-v2-finetuned-imdb")
# Run inference
sentences = [
    ' "Oliver Twist" (1985)  In a storm, in a workhouse, to a nameless woman, young Oliver Twist is born into parish care where he\'s overworked and underfed. As he grows older his adventures take him from the countryside to London, through harsh treatment, kindness, an undertaker, and a thieves\' dens, where he makes friends and enemies. But all the time he is pursued by the mysterious Monks, who hires Fagin to turn Oliver into a thief. Oliver is rescued by chance and kind friends. But it\'s a puzzle of legitimacy, inheritance, and identity that Oliver\'s friends must attempt to unravel before Monks can destroy Oliver.',
    'drama',
    'documentary',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 384]

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

Evaluation

Metrics

Binary Classification

Evaluated with BinaryClassificationEvaluator

Metric	Value
cosine_accuracy	0.9007
cosine_accuracy_threshold	0.602
cosine_f1	0.4643
cosine_f1_threshold	0.5201
cosine_precision	0.4201
cosine_recall	0.5189
cosine_ap	0.4637
dot_accuracy	0.9007
dot_accuracy_threshold	0.602
dot_f1	0.4643
dot_f1_threshold	0.5201
dot_precision	0.4201
dot_recall	0.5189
dot_ap	0.4637
manhattan_accuracy	0.9003
manhattan_accuracy_threshold	13.5474
manhattan_f1	0.4582
manhattan_f1_threshold	15.1497
manhattan_precision	0.4095
manhattan_recall	0.52
manhattan_ap	0.4579
euclidean_accuracy	0.9007
euclidean_accuracy_threshold	0.8922
euclidean_f1	0.4643
euclidean_f1_threshold	0.9797
euclidean_precision	0.4201
euclidean_recall	0.5189
euclidean_ap	0.4637
max_accuracy	0.9007
max_accuracy_threshold	13.5474
max_f1	0.4643
max_f1_threshold	15.1497
max_precision	0.4201
max_recall	0.52
max_ap	0.4637

Triplet

Evaluated with TripletEvaluator

Metric	Value
cosine_accuracy	0.6382
dot_accuracy	0.3618
manhattan_accuracy	0.6227
euclidean_accuracy	0.6382
max_accuracy	0.6382

Training Details

Training Dataset

imdb-triplet

Dataset: imdb-triplet
Size: 43,371 training samples
Columns: anchor and positive
Approximate statistics based on the first 1000 samples:
anchor positive
type string string
details
min: 31 tokens
mean: 129.65 tokens
max: 256 tokens
min: 3 tokens
mean: 3.0 tokens
max: 3 tokens

Samples:

anchor	positive
`A Metafísica dos Chocolates (1967) Beautiful girls (pre-teens, adolescents, and young women) in street scenes and one of them visiting a chocolate factory, where all the workers are young women, too. A poetic text and an extract from a major Portuguese poet, convey to us the sensual feeling of choosing, unwrapping, and munching chocolate.`	`short`
`Thai Jashe! (2016) Thai Jashe! is an upcoming Gujarati film written and directed by Nirav Barot. It is about the struggles of a middle class man to achieve his goals in the metro-city Ahmedabad. The film stars Manoj Joshi, Malhar Thakar and Monal Gajjar.`	`drama`
`Vuelco (2005) A teenage boy rides out of town to meet a a girl in the countryside. She is deaf, and he explains the different means he uses to get her attention when she has not seen him. Then they say goodbye, with one poignant hug and a desperate yell punctuating their final farewell.`	`short`

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

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: steps
per_device_train_batch_size: 256
per_device_eval_batch_size: 256
num_train_epochs: 5
warmup_ratio: 0.1
fp16: True
batch_sampler: no_duplicates
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: 256
per_device_eval_batch_size: 256
per_gpu_train_batch_size: None
per_gpu_eval_batch_size: None
gradient_accumulation_steps: 1
eval_accumulation_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: 5
max_steps: -1
lr_scheduler_type: linear
lr_scheduler_kwargs: {}
warmup_ratio: 0.1
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}
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: False
hub_always_push: False
gradient_checkpointing: False
gradient_checkpointing_kwargs: None
include_inputs_for_metrics: False
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
dispatch_batches: None
split_batches: None
include_tokens_per_second: False
include_num_input_tokens_seen: False
neftune_noise_alpha: None
optim_target_modules: None
batch_eval_metrics: False
batch_sampler: no_duplicates
multi_dataset_batch_sampler: round_robin

Training Logs

Epoch	Step	Training Loss	max_accuracy	max_ap
0	0	-	0.6382	0.2004
0.5882	100	1.7867	-	0.3542
1.1765	200	1.3073	-	0.4564
1.7647	300	1.266	-	0.3862
2.3529	400	1.1889	-	0.4011
2.9412	500	1.1554	-	0.4398
3.5294	600	1.1558	-	0.4386
4.1176	700	1.1555	-	0.4566
4.7059	800	1.0835	-	0.4637

Framework Versions

Python: 3.10.13
Sentence Transformers: 3.0.1
Transformers: 4.41.2
PyTorch: 2.1.2
Accelerate: 0.30.1
Datasets: 2.19.2
Tokenizers: 0.19.1

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

bibtex

@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}
}

	anchor	positive
type	string	string
details	min: 31 tokens mean: 129.65 tokens max: 256 tokens	min: 3 tokens mean: 3.0 tokens max: 3 tokens