metadata

base_model: sentence-transformers/all-MiniLM-L12-v2
library_name: sentence-transformers
metrics:
  - pearson_cosine
  - spearman_cosine
  - pearson_manhattan
  - spearman_manhattan
  - pearson_euclidean
  - spearman_euclidean
  - pearson_dot
  - spearman_dot
  - pearson_max
  - spearman_max
pipeline_tag: sentence-similarity
tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:100000
  - loss:CosineSimilarityLoss
widget:
  - source_sentence: Face off with a ref mid-hockey game in an arena.
    sentences:
      - Nobody is playing
      - >-
        A mustached man in a patterned shirt watches a boat painted blue and
        orange.
      - Two adults makes calls on there cell phones during there lunch breaks.
  - source_sentence: >-
      A group of people, one holding a yellow and blue umbrella, are standing at
      the top of some stairs.
    sentences:
      - One person wields an umbrella.
      - A girl is on the beach.
      - A man is on his couch.
  - source_sentence: >-
      A man waiting for the results of the machine after doing an experiment in
      his laboratory.
    sentences:
      - There is a man playing an instrument while running
      - A man in a lab waits to get more information about his experiment.
      - The graffiti artists admire their work.
  - source_sentence: People in a tent shelter near the bottom of stairs.
    sentences:
      - A boy has fallen asleep during dinner.
      - Three men address a crowd.
      - People are in a makeshift shelter at the foot of a staircase.
  - source_sentence: A female researcher looking through a microscope.
    sentences:
      - A man misses the rope and falls
      - A small girl is playing video games
      - A woman is researching with a microscope.
model-index:
  - name: SentenceTransformer based on sentence-transformers/all-MiniLM-L12-v2
    results:
      - task:
          type: semantic-similarity
          name: Semantic Similarity
        dataset:
          name: snli dev
          type: snli-dev
        metrics:
          - type: pearson_cosine
            value: 0.48994508338253345
            name: Pearson Cosine
          - type: spearman_cosine
            value: 0.4778683474663533
            name: Spearman Cosine
          - type: pearson_manhattan
            value: 0.46917600703738915
            name: Pearson Manhattan
          - type: spearman_manhattan
            value: 0.47754796729416876
            name: Spearman Manhattan
          - type: pearson_euclidean
            value: 0.46924620767742137
            name: Pearson Euclidean
          - type: spearman_euclidean
            value: 0.4778683474663533
            name: Spearman Euclidean
          - type: pearson_dot
            value: 0.48994508631435785
            name: Pearson Dot
          - type: spearman_dot
            value: 0.4778683472855999
            name: Spearman Dot
          - type: pearson_max
            value: 0.48994508631435785
            name: Pearson Max
          - type: spearman_max
            value: 0.4778683474663533
            name: Spearman Max

SentenceTransformer based on sentence-transformers/all-MiniLM-L12-v2

This is a sentence-transformers model finetuned from sentence-transformers/all-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/all-MiniLM-L12-v2
Maximum Sequence Length: 128 tokens
Output Dimensionality: 384 tokens
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': 128, '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.

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("Nessrine9/finetuned2-snli-MiniLM-L12-v2")
# Run inference
sentences = [
    'A female researcher looking through a microscope.',
    'A woman is researching with a microscope.',
    'A small girl is playing video games',
]
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

Semantic Similarity

Dataset: snli-dev
Evaluated with EmbeddingSimilarityEvaluator

Metric	Value
pearson_cosine	0.4899
spearman_cosine	0.4779
pearson_manhattan	0.4692
spearman_manhattan	0.4775
pearson_euclidean	0.4692
spearman_euclidean	0.4779
pearson_dot	0.4899
spearman_dot	0.4779
pearson_max	0.4899
spearman_max	0.4779

Training Details

Training Dataset

Unnamed Dataset

Size: 100,000 training samples
Columns: sentence_0, sentence_1, and label
Approximate statistics based on the first 1000 samples:
sentence_0 sentence_1 label
type string string float
details
min: 7 tokens
mean: 16.32 tokens
max: 86 tokens

min: 4 tokens
mean: 10.46 tokens
max: 32 tokens

min: 0.0
mean: 0.5
max: 1.0

	sentence_0	sentence_1	label
type	string	string	float
details	min: 7 tokens mean: 16.32 tokens max: 86 tokens	min: 4 tokens mean: 10.46 tokens max: 32 tokens	min: 0.0 mean: 0.5 max: 1.0

Samples:

sentence_0	sentence_1	label
`A man wearing jeans and a t-shirt plays guitar for a smiling woman and child as they sit on a staircase near red and orange balloons.`	`A man is in jail.`	`1.0`
`A boy wearing blue short standing on the traffic signal pole.`	`The boy is carrying his school books.`	`0.5`
`Several people on a busy street or perhaps at a fair.`	`They are walkng.`	`0.5`

Loss: CosineSimilarityLoss with these parameters:

{
    "loss_fct": "torch.nn.modules.loss.MSELoss"
}

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: steps
per_device_train_batch_size: 16
per_device_eval_batch_size: 16
num_train_epochs: 4
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: 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: 4
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}
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
eval_on_start: False
eval_use_gather_object: False
batch_sampler: batch_sampler
multi_dataset_batch_sampler: round_robin

Training Logs

Epoch	Step	Training Loss	snli-dev_spearman_max
0.08	500	0.1832	0.3114
0.16	1000	0.1489	0.3518
0.24	1500	0.1468	0.3697
0.32	2000	0.1411	0.3723
0.4	2500	0.14	0.4062
0.48	3000	0.1366	0.3923
0.56	3500	0.1379	0.4143
0.64	4000	0.1357	0.3928
0.72	4500	0.1331	0.4067
0.8	5000	0.1338	0.4293
0.88	5500	0.1294	0.4183
0.96	6000	0.1305	0.4402
1.0	6250	-	0.4454
1.04	6500	0.1303	0.4408
1.12	7000	0.1275	0.4416
1.2	7500	0.1285	0.4287
1.28	8000	0.125	0.4404
1.3600	8500	0.1253	0.4408
1.44	9000	0.1246	0.4293
1.52	9500	0.126	0.4535
1.6	10000	0.1257	0.4455
1.6800	10500	0.1264	0.4520
1.76	11000	0.1248	0.4526
1.8400	11500	0.1208	0.4631
1.92	12000	0.1236	0.4635
2.0	12500	0.1239	0.4573
2.08	13000	0.1209	0.4569
2.16	13500	0.1194	0.4642
2.24	14000	0.1206	0.4539
2.32	14500	0.117	0.4633
2.4	15000	0.1171	0.4657
2.48	15500	0.1181	0.4633
2.56	16000	0.1197	0.4552
2.64	16500	0.1182	0.4670
2.7200	17000	0.1155	0.4684
2.8	17500	0.1171	0.4640
2.88	18000	0.1139	0.4715
2.96	18500	0.1164	0.4769
3.0	18750	-	0.4709
3.04	19000	0.1151	0.4704
3.12	19500	0.1144	0.4759
3.2	20000	0.1121	0.4795
3.2800	20500	0.1104	0.4697
3.36	21000	0.1127	0.4763
3.44	21500	0.1115	0.4742
3.52	22000	0.1126	0.4697
3.6	22500	0.1123	0.4735
3.68	23000	0.1132	0.4750
3.76	23500	0.1127	0.4743
3.84	24000	0.1086	0.4752
3.92	24500	0.1107	0.4781
4.0	25000	0.1114	0.4779

Framework Versions

Python: 3.10.12
Sentence Transformers: 3.2.1
Transformers: 4.44.2
PyTorch: 2.5.0+cu121
Accelerate: 0.34.2
Datasets: 3.0.2
Tokenizers: 0.19.1

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",
}