Omartificial-Intelligence-Space's picture

update readme.md

fa55cb1 verified about 1 year ago

39.6 kB

language: []
library_name: sentence-transformers
tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:557850
  - loss:MatryoshkaLoss
  - loss:MultipleNegativesRankingLoss
base_model: intfloat/multilingual-e5-small
datasets: []
metrics:
  - pearson_cosine
  - spearman_cosine
  - pearson_manhattan
  - spearman_manhattan
  - pearson_euclidean
  - spearman_euclidean
  - pearson_dot
  - spearman_dot
  - pearson_max
  - spearman_max
widget:
  - source_sentence: ذكر متوازن بعناية يقف على قدم واحدة بالقرب من منطقة شاطئ المحيط النظيفة
    sentences:
      - رجل يقدم عرضاً
      - هناك رجل بالخارج قرب الشاطئ
      - رجل يجلس على أريكه
  - source_sentence: رجل يقفز إلى سريره القذر
    sentences:
      - السرير قذر.
      - رجل يضحك أثناء غسيل الملابس
      - الرجل على القمر
  - source_sentence: الفتيات بالخارج
    sentences:
      - امرأة تلف الخيط إلى كرات بجانب كومة من الكرات
      - فتيان يركبان في جولة متعة
      - >-
        ثلاث فتيات يقفون سوية في غرفة واحدة تستمع وواحدة تكتب على الحائط
        والثالثة تتحدث إليهن
  - source_sentence: الرجل يرتدي قميصاً أزرق.
    sentences:
      - >-
        رجل يرتدي قميصاً أزرق يميل إلى الجدار بجانب الطريق مع شاحنة زرقاء وسيارة
        حمراء مع الماء في الخلفية.
      - كتاب القصص مفتوح
      - رجل يرتدي قميص أسود يعزف على الجيتار.
  - source_sentence: يجلس شاب ذو شعر أشقر على الحائط يقرأ جريدة بينما تمر امرأة وفتاة شابة.
    sentences:
      - ذكر شاب ينظر إلى جريدة بينما تمر إمرأتان بجانبه
      - رجل يستلقي على وجهه على مقعد في الحديقة.
      - الشاب نائم بينما الأم تقود ابنتها إلى الحديقة
pipeline_tag: sentence-similarity
model-index:
  - name: SentenceTransformer based on intfloat/multilingual-e5-small
    results:
      - dataset:
          config: ar
          name: MTEB MIRACLRetrievalHardNegatives (ar)
          revision: 95c8db7d4a6e9c1d8a60601afd63d553ae20a2eb
          split: dev
          type: mteb/miracl-hard-negatives
        metrics:
          - type: main_score
            value: 33.441
        task:
          type: Retrieval
      - dataset:
          config: ara-ara
          name: MTEB MLQARetrieval (ara-ara)
          revision: 397ed406c1a7902140303e7faf60fff35b58d285
          split: test
          type: facebook/mlqa
        metrics:
          - type: main_score
            value: 64.488
        task:
          type: Retrieval
      - dataset:
          config: ar
          name: MTEB MintakaRetrieval (ar)
          revision: efa78cc2f74bbcd21eff2261f9e13aebe40b814e
          split: test
          type: jinaai/mintakaqa
        metrics:
          - type: main_score
            value: 16.162
        task:
          type: Retrieval
      - dataset:
          config: default
          name: MTEB SadeemQuestionRetrieval (default)
          revision: 3cb0752b182e5d5d740df547748b06663c8e0bd9
          split: test
          type: sadeem-ai/sadeem-ar-eval-retrieval-questions
        metrics:
          - type: main_score
            value: 63.235
        task:
          type: Retrieval
      - task:
          type: semantic-similarity
          name: Semantic Similarity
        dataset:
          name: sts test 384
          type: sts-test-384
        metrics:
          - type: pearson_cosine
            value: 0.7883137447514015
            name: Pearson Cosine
          - type: spearman_cosine
            value: 0.7971624317482785
            name: Spearman Cosine
          - type: pearson_manhattan
            value: 0.7845904338398069
            name: Pearson Manhattan
          - type: spearman_manhattan
            value: 0.7939541836133244
            name: Spearman Manhattan
          - type: pearson_euclidean
            value: 0.7882887522003604
            name: Pearson Euclidean
          - type: spearman_euclidean
            value: 0.7971601260546269
            name: Spearman Euclidean
          - type: pearson_dot
            value: 0.7883137483129774
            name: Pearson Dot
          - type: spearman_dot
            value: 0.7971605875966696
            name: Spearman Dot
          - type: pearson_max
            value: 0.7883137483129774
            name: Pearson Max
          - type: spearman_max
            value: 0.7971624317482785
            name: Spearman Max
      - task:
          type: semantic-similarity
          name: Semantic Similarity
        dataset:
          name: sts test 256
          type: sts-test-256
        metrics:
          - type: pearson_cosine
            value: 0.7851969391652749
            name: Pearson Cosine
          - type: spearman_cosine
            value: 0.7968026743946358
            name: Spearman Cosine
          - type: pearson_manhattan
            value: 0.7852783784725356
            name: Pearson Manhattan
          - type: spearman_manhattan
            value: 0.7935883492889713
            name: Spearman Manhattan
          - type: pearson_euclidean
            value: 0.7882018230746569
            name: Pearson Euclidean
          - type: spearman_euclidean
            value: 0.7963116553267949
            name: Spearman Euclidean
          - type: pearson_dot
            value: 0.7786421988393841
            name: Pearson Dot
          - type: spearman_dot
            value: 0.7867782644180616
            name: Spearman Dot
          - type: pearson_max
            value: 0.7882018230746569
            name: Pearson Max
          - type: spearman_max
            value: 0.7968026743946358
            name: Spearman Max
      - task:
          type: semantic-similarity
          name: Semantic Similarity
        dataset:
          name: sts test 128
          type: sts-test-128
        metrics:
          - type: pearson_cosine
            value: 0.7754967709350954
            name: Pearson Cosine
          - type: spearman_cosine
            value: 0.7933453885370457
            name: Spearman Cosine
          - type: pearson_manhattan
            value: 0.7832834632297865
            name: Pearson Manhattan
          - type: spearman_manhattan
            value: 0.7907589269176767
            name: Spearman Manhattan
          - type: pearson_euclidean
            value: 0.7867583047946054
            name: Pearson Euclidean
          - type: spearman_euclidean
            value: 0.7935816990844704
            name: Spearman Euclidean
          - type: pearson_dot
            value: 0.7317253736607925
            name: Pearson Dot
          - type: spearman_dot
            value: 0.7335574962775742
            name: Spearman Dot
          - type: pearson_max
            value: 0.7867583047946054
            name: Pearson Max
          - type: spearman_max
            value: 0.7935816990844704
            name: Spearman Max
      - task:
          type: semantic-similarity
          name: Semantic Similarity
        dataset:
          name: sts test 64
          type: sts-test-64
        metrics:
          - type: pearson_cosine
            value: 0.7625204599039478
            name: Pearson Cosine
          - type: spearman_cosine
            value: 0.7837078735068292
            name: Spearman Cosine
          - type: pearson_manhattan
            value: 0.7752889433866854
            name: Pearson Manhattan
          - type: spearman_manhattan
            value: 0.7790888579029828
            name: Spearman Manhattan
          - type: pearson_euclidean
            value: 0.777961287133872
            name: Pearson Euclidean
          - type: spearman_euclidean
            value: 0.7815940757356076
            name: Spearman Euclidean
          - type: pearson_dot
            value: 0.6685094830550401
            name: Pearson Dot
          - type: spearman_dot
            value: 0.6621206899696827
            name: Spearman Dot
          - type: pearson_max
            value: 0.777961287133872
            name: Pearson Max
          - type: spearman_max
            value: 0.7837078735068292
            name: Spearman Max

SentenceTransformer based on intfloat/multilingual-e5-small

This is a sentence-transformers model finetuned from intfloat/multilingual-e5-small on the Omartificial-Intelligence-Space/arabic-n_li-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: intfloat/multilingual-e5-small
Maximum Sequence Length: 512 tokens
Output Dimensionality: 384 tokens
Similarity Function: Cosine Similarity
Training Dataset:
- Omartificial-Intelligence-Space/arabic-n_li-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': 512, '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("Omartificial-Intelligence-Space/E5-Matro")
# Run inference
sentences = [
    'يجلس شاب ذو شعر أشقر على الحائط يقرأ جريدة بينما تمر امرأة وفتاة شابة.',
    'ذكر شاب ينظر إلى جريدة بينما تمر إمرأتان بجانبه',
    'الشاب نائم بينما الأم تقود ابنتها إلى الحديقة',
]
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: sts-test-384
Evaluated with EmbeddingSimilarityEvaluator

Metric	Value
pearson_cosine	0.7883
spearman_cosine	0.7972
pearson_manhattan	0.7846
spearman_manhattan	0.794
pearson_euclidean	0.7883
spearman_euclidean	0.7972
pearson_dot	0.7883
spearman_dot	0.7972
pearson_max	0.7883
spearman_max	0.7972

Semantic Similarity

Dataset: sts-test-256
Evaluated with EmbeddingSimilarityEvaluator

Metric	Value
pearson_cosine	0.7852
spearman_cosine	0.7968
pearson_manhattan	0.7853
spearman_manhattan	0.7936
pearson_euclidean	0.7882
spearman_euclidean	0.7963
pearson_dot	0.7786
spearman_dot	0.7868
pearson_max	0.7882
spearman_max	0.7968

Semantic Similarity

Dataset: sts-test-128
Evaluated with EmbeddingSimilarityEvaluator

Metric	Value
pearson_cosine	0.7755
spearman_cosine	0.7933
pearson_manhattan	0.7833
spearman_manhattan	0.7908
pearson_euclidean	0.7868
spearman_euclidean	0.7936
pearson_dot	0.7317
spearman_dot	0.7336
pearson_max	0.7868
spearman_max	0.7936

Semantic Similarity

Dataset: sts-test-64
Evaluated with EmbeddingSimilarityEvaluator

Metric	Value
pearson_cosine	0.7625
spearman_cosine	0.7837
pearson_manhattan	0.7753
spearman_manhattan	0.7791
pearson_euclidean	0.778
spearman_euclidean	0.7816
pearson_dot	0.6685
spearman_dot	0.6621
pearson_max	0.778
spearman_max	0.7837

Training Details

Training Dataset

Omartificial-Intelligence-Space/arabic-n_li-triplet

Dataset: Omartificial-Intelligence-Space/arabic-n_li-triplet
Size: 557,850 training samples
Columns: anchor, positive, and negative

Approximate statistics based on the first 1000 samples:

	anchor	positive	negative
type	string	string	string
details	min: 5 tokens mean: 10.33 tokens max: 52 tokens	min: 5 tokens mean: 13.21 tokens max: 49 tokens	min: 5 tokens mean: 15.32 tokens max: 53 tokens

Samples:

anchor	positive	negative
`شخص على حصان يقفز فوق طائرة معطلة`	`شخص في الهواء الطلق، على حصان.`	`شخص في مطعم، يطلب عجة.`
`أطفال يبتسمون و يلوحون للكاميرا`	`هناك أطفال حاضرون`	`الاطفال يتجهمون`
`صبي يقفز على لوح التزلج في منتصف الجسر الأحمر.`	`الفتى يقوم بخدعة التزلج`	`الصبي يتزلج على الرصيف`

Loss: MatryoshkaLoss with these parameters:

{
    "loss": "MultipleNegativesRankingLoss",
    "matryoshka_dims": [
        384,
        256,
        128,
        64
    ],
    "matryoshka_weights": [
        1,
        1,
        1,
        1
    ],
    "n_dims_per_step": -1
}

Evaluation Dataset

Omartificial-Intelligence-Space/arabic-n_li-triplet

Dataset: Omartificial-Intelligence-Space/arabic-n_li-triplet
Size: 6,584 evaluation samples
Columns: anchor, positive, and negative

Approximate statistics based on the first 1000 samples:

	anchor	positive	negative
type	string	string	string
details	min: 5 tokens mean: 21.86 tokens max: 105 tokens	min: 4 tokens mean: 10.22 tokens max: 49 tokens	min: 4 tokens mean: 11.2 tokens max: 33 tokens

Samples:

anchor	positive	negative
`امرأتان يتعانقان بينما يحملان حزمة`	`إمرأتان يحملان حزمة`	`الرجال يتشاجرون خارج مطعم`
`طفلين صغيرين يرتديان قميصاً أزرق، أحدهما يرتدي الرقم 9 والآخر يرتدي الرقم 2 يقفان على خطوات خشبية في الحمام ويغسلان أيديهما في المغسلة.`	`طفلين يرتديان قميصاً مرقماً يغسلون أيديهم`	`طفلين يرتديان سترة يذهبان إلى المدرسة`
`رجل يبيع الدونات لعميل خلال معرض عالمي أقيم في مدينة أنجليس`	`رجل يبيع الدونات لعميل`	`امرأة تشرب قهوتها في مقهى صغير`

Loss: MatryoshkaLoss with these parameters:

{
    "loss": "MultipleNegativesRankingLoss",
    "matryoshka_dims": [
        384,
        256,
        128,
        64
    ],
    "matryoshka_weights": [
        1,
        1,
        1,
        1
    ],
    "n_dims_per_step": -1
}

Training Hyperparameters

Non-Default Hyperparameters

per_device_train_batch_size: 32
per_device_eval_batch_size: 32
warmup_ratio: 0.1
fp16: True
batch_sampler: no_duplicates

All Hyperparameters

Click to expand

overwrite_output_dir: False
do_predict: False
prediction_loss_only: True
per_device_train_batch_size: 32
per_device_eval_batch_size: 32
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: 3
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
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, '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_sampler: no_duplicates
multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	sts-test-128_spearman_cosine	sts-test-256_spearman_cosine	sts-test-384_spearman_cosine	sts-test-64_spearman_cosine
0.0344	200	13.1208	-	-	-	-
0.0688	400	9.1894	-	-	-	-
0.1033	600	8.0222	-	-	-	-
0.1377	800	7.2405	-	-	-	-
0.1721	1000	7.1622	-	-	-	-
0.2065	1200	6.4282	-	-	-	-
0.2409	1400	6.0936	-	-	-	-
0.2753	1600	5.99	-	-	-	-
0.3098	1800	5.6939	-	-	-	-
0.3442	2000	5.694	-	-	-	-
0.3786	2200	5.2366	-	-	-	-
0.4130	2400	5.2994	-	-	-	-
0.4474	2600	5.2079	-	-	-	-
0.4818	2800	5.0532	-	-	-	-
0.5163	3000	4.9978	-	-	-	-
0.5507	3200	5.1764	-	-	-	-
0.5851	3400	5.1315	-	-	-	-
0.6195	3600	5.0198	-	-	-	-
0.6539	3800	5.0308	-	-	-	-
0.6883	4000	5.1631	-	-	-	-
0.7228	4200	4.7916	-	-	-	-
0.7572	4400	4.363	-	-	-	-
0.7916	4600	3.2357	-	-	-	-
0.8260	4800	2.9915	-	-	-	-
0.8604	5000	2.8143	-	-	-	-
0.8949	5200	2.6125	-	-	-	-
0.9293	5400	2.5493	-	-	-	-
0.9637	5600	2.4991	-	-	-	-
0.9981	5800	2.163	-	-	-	-
1.0325	6000	0.0	-	-	-	-
1.0669	6200	0.0	-	-	-	-
1.1014	6400	0.0	-	-	-	-
1.1358	6600	0.0	-	-	-	-
1.1702	6800	0.0	-	-	-	-
1.2046	7000	0.0	-	-	-	-
1.2390	7200	0.0	-	-	-	-
1.2734	7400	0.0	-	-	-	-
1.3079	7600	0.0	-	-	-	-
1.3423	7800	0.0	-	-	-	-
1.3767	8000	0.0	-	-	-	-
1.4111	8200	0.0037	-	-	-	-
1.4455	8400	0.0372	-	-	-	-
1.4800	8600	0.0221	-	-	-	-
1.0229	8800	4.3738	-	-	-	-
1.0573	9000	6.338	-	-	-	-
1.0917	9200	6.2223	-	-	-	-
1.1261	9400	5.8673	-	-	-	-
1.1606	9600	5.5907	-	-	-	-
1.1950	9800	5.0307	-	-	-	-
1.2294	10000	4.9193	-	-	-	-
1.2638	10200	4.8798	-	-	-	-
1.2982	10400	4.401	-	-	-	-
1.3326	10600	4.2705	-	-	-	-
1.3671	10800	4.3023	-	-	-	-
1.4015	11000	4.1344	-	-	-	-
1.4359	11200	4.0464	-	-	-	-
1.4703	11400	4.0115	-	-	-	-
1.5047	11600	3.9206	-	-	-	-
1.5391	11800	4.0106	-	-	-	-
1.5736	12000	4.1365	-	-	-	-
1.6080	12200	4.0401	-	-	-	-
1.6424	12400	4.0602	-	-	-	-
1.6768	12600	4.076	-	-	-	-
1.7112	12800	3.97	-	-	-	-
1.7457	13000	3.7905	-	-	-	-
1.7801	13200	2.414	-	-	-	-
1.8145	13400	2.1811	-	-	-	-
1.8489	13600	2.1183	-	-	-	-
1.8833	13800	2.0578	-	-	-	-
1.9177	14000	2.0173	-	-	-	-
1.9522	14200	2.0093	-	-	-	-
1.9866	14400	1.9467	-	-	-	-
2.0210	14600	0.4674	-	-	-	-
2.0554	14800	0.0	-	-	-	-
2.0898	15000	0.0	-	-	-	-
2.1242	15200	0.0	-	-	-	-
2.1587	15400	0.0	-	-	-	-
2.1931	15600	0.0	-	-	-	-
2.2275	15800	0.0	-	-	-	-
2.2619	16000	0.0	-	-	-	-
2.2963	16200	0.0	-	-	-	-
2.3308	16400	0.0	-	-	-	-
2.3652	16600	0.0	-	-	-	-
2.3996	16800	0.0	-	-	-	-
2.4340	17000	0.0	-	-	-	-
2.4684	17200	0.0256	-	-	-	-
2.0114	17400	2.4155	-	-	-	-
2.0170	17433	-	0.7933	0.7968	0.7972	0.7837

Framework Versions

Python: 3.9.18
Sentence Transformers: 3.0.1
Transformers: 4.40.0
PyTorch: 2.2.2+cu121
Accelerate: 0.26.1
Datasets: 2.19.0
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",
}

MatryoshkaLoss

@misc{kusupati2024matryoshka,
    title={Matryoshka Representation Learning}, 
    author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi},
    year={2024},
    eprint={2205.13147},
    archivePrefix={arXiv},
    primaryClass={cs.LG}
}

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}
}