SentenceTransformer based on axiomepic/maux-gte-persian-v3-finetuned-bce

This is a sentence-transformers model finetuned from axiomepic/maux-gte-persian-v3-finetuned-bce. 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: axiomepic/maux-gte-persian-v3-finetuned-bce
• Maximum Sequence Length: 8192 tokens
• Output Dimensionality: 768 dimensions
• Similarity Function: Dot Product

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': 8192, 'do_lower_case': False, 'architecture': 'NewModel'})
  (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, '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("axiomepic/maux-gte-persian-v3-finetuned-bce")
# Run inference
sentences = [
    'توالت فرنگی کوچک',
    'نشستن روی توالت فرنگی',
    'leg with dumbbells',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities)
# tensor([[0.7279, 0.7355, 0.2055],
#         [0.7355, 0.7768, 0.2210],
#         [0.2055, 0.2210, 2.5745]])

Evaluation

Metrics

Binary Classification

• Dataset: seo-bce-eval
• Evaluated with BinaryClassificationEvaluator

Metric	Value
dot_accuracy	0.6364
dot_accuracy_threshold	1.2505
dot_f1	0.5554
dot_f1_threshold	-0.1131
dot_precision	0.3883
dot_recall	0.9748
dot_ap	0.439
dot_mcc	0.1287

Training Details

Training Dataset

Unnamed Dataset

• Size: 52,893 training samples
• Columns: sentence1, sentence2, and label
• Approximate statistics based on the first 1000 samples:

  	sentence1	sentence2	label
  type	string	string	float
  details	min: 3 tokens mean: 6.59 tokens max: 18 tokens	min: 3 tokens mean: 6.27 tokens max: 17 tokens	min: 0.0 mean: 0.38 max: 1.0

• Samples:

  sentence1	sentence2	label
  جک سقفی	کامیونت جک	1.0
  سفارت فرانسه در تهران	dutch embassy paris	0.0
  بهترین آبمیوه گیری	فواید آب	0.0

• Loss: main.DotProductBCELoss

Evaluation Dataset

Unnamed Dataset

• Size: 5,877 evaluation samples
• Columns: sentence1, sentence2, and label
• Approximate statistics based on the first 1000 samples:

  	sentence1	sentence2	label
  type	string	string	float
  details	min: 3 tokens mean: 6.55 tokens max: 15 tokens	min: 3 tokens mean: 6.24 tokens max: 18 tokens	min: 0.0 mean: 0.38 max: 1.0

• Samples:

  sentence1	sentence2	label
  ظروف آلومینیومی یکبار مصرف	ظروف پلاستیکی	1.0
  اخبار مهاجرین	current events immigration	1.0
  معامله آسان	buy sale trade	1.0

• Loss: main.DotProductBCELoss

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: epoch
• per_device_train_batch_size: 16
• per_device_eval_batch_size: 64
• learning_rate: 2e-05
• num_train_epochs: 4
• warmup_ratio: 0.1
• load_best_model_at_end: True

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: epoch
• prediction_loss_only: True
• per_device_train_batch_size: 16
• per_device_eval_batch_size: 64
• 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: 2e-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: 4
• 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
• 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: True
• 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: proportional
• router_mapping: {}
• learning_rate_mapping: {}

Training Logs

Click to expand

Epoch	Step	Training Loss	Validation Loss	seo-bce-eval_dot_ap
0.0030	10	114.2302	-	-
0.0060	20	113.3976	-	-
0.0091	30	115.9475	-	-
0.0121	40	122.7224	-	-
0.0151	50	112.9675	-	-
0.0181	60	114.5466	-	-
0.0212	70	116.0311	-	-
0.0242	80	99.6088	-	-
0.0272	90	84.4752	-	-
0.0302	100	82.4359	-	-
0.0333	110	77.9756	-	-
0.0363	120	70.7338	-	-
0.0393	130	61.25	-	-
0.0423	140	37.6046	-	-
0.0454	150	34.1171	-	-
0.0484	160	30.6499	-	-
0.0514	170	22.5837	-	-
0.0544	180	17.1006	-	-
0.0575	190	12.3175	-	-
0.0605	200	8.4437	-	-
0.0635	210	5.3229	-	-
0.0665	220	3.8265	-	-
0.0696	230	3.1104	-	-
0.0726	240	2.2156	-	-
0.0756	250	1.8534	-	-
0.0786	260	1.4878	-	-
0.0817	270	1.5639	-	-
0.0847	280	1.375	-	-
0.0877	290	1.2778	-	-
0.0907	300	1.3444	-	-
0.0938	310	1.4949	-	-
0.0968	320	1.3558	-	-
0.0998	330	1.0515	-	-
0.1028	340	1.1467	-	-
0.1059	350	1.0924	-	-
0.1089	360	1.2178	-	-
0.1119	370	1.2049	-	-
0.1149	380	1.1147	-	-
0.1180	390	0.9555	-	-
0.1210	400	1.2939	-	-
0.1240	410	1.1083	-	-
0.1270	420	1.0736	-	-
0.1301	430	1.3217	-	-
0.1331	440	1.138	-	-
0.1361	450	1.1887	-	-
0.1391	460	1.0804	-	-
0.1422	470	1.0579	-	-
0.1452	480	0.9958	-	-
0.1482	490	1.0312	-	-
0.1512	500	1.0218	-	-
0.1543	510	1.0107	-	-
0.1573	520	1.0576	-	-
0.1603	530	1.0445	-	-
0.1633	540	0.9355	-	-
0.1664	550	0.946	-	-
0.1694	560	0.9423	-	-
0.1724	570	1.0467	-	-
0.1754	580	0.9928	-	-
0.1785	590	0.9504	-	-
0.1815	600	0.9529	-	-
0.1845	610	1.0225	-	-
0.1875	620	0.9391	-	-
0.1906	630	0.9984	-	-
0.1936	640	0.965	-	-
0.1966	650	0.93	-	-
0.1996	660	0.9429	-	-
0.2027	670	0.9713	-	-
0.2057	680	1.0207	-	-
0.2087	690	0.8401	-	-
0.2117	700	0.9345	-	-
0.2148	710	0.9107	-	-
0.2178	720	0.9833	-	-
0.2208	730	0.927	-	-
0.2238	740	0.9652	-	-
0.2269	750	0.9116	-	-
0.2299	760	0.8711	-	-
0.2329	770	0.9002	-	-
0.2359	780	0.9852	-	-
0.2390	790	1.0248	-	-
0.2420	800	0.779	-	-
0.2450	810	0.8977	-	-
0.2480	820	0.9285	-	-
0.2511	830	0.9813	-	-
0.2541	840	0.9001	-	-
0.2571	850	0.9476	-	-
0.2601	860	0.934	-	-
0.2632	870	0.8883	-	-
0.2662	880	1.0017	-	-
0.2692	890	0.9306	-	-
0.2722	900	0.904	-	-
0.2753	910	0.8299	-	-
0.2783	920	0.8661	-	-
0.2813	930	0.9472	-	-
0.2843	940	0.8801	-	-
0.2874	950	1.0444	-	-
0.2904	960	0.9591	-	-
0.2934	970	0.864	-	-
0.2964	980	0.9015	-	-
0.2995	990	0.849	-	-
0.3025	1000	0.8284	-	-
0.3055	1010	0.9114	-	-
0.3085	1020	0.9288	-	-
0.3116	1030	0.8625	-	-
0.3146	1040	0.9262	-	-
0.3176	1050	0.8887	-	-
0.3206	1060	0.9522	-	-
0.3237	1070	0.8593	-	-
0.3267	1080	0.8576	-	-
0.3297	1090	0.8723	-	-
0.3327	1100	0.9361	-	-
0.3358	1110	0.8741	-	-
0.3388	1120	0.8795	-	-
0.3418	1130	0.915	-	-
0.3448	1140	0.9236	-	-
0.3479	1150	0.8555	-	-
0.3509	1160	0.9535	-	-
0.3539	1170	0.8852	-	-
0.3569	1180	0.9201	-	-
0.3600	1190	0.8791	-	-
0.3630	1200	0.8594	-	-
0.3660	1210	0.8514	-	-
0.3690	1220	0.9417	-	-
0.3721	1230	0.8887	-	-
0.3751	1240	0.9052	-	-
0.3781	1250	0.8686	-	-
0.3811	1260	0.8952	-	-
0.3842	1270	0.8843	-	-
0.3872	1280	0.8415	-	-
0.3902	1290	0.8904	-	-
0.3932	1300	0.9342	-	-
0.3962	1310	0.9093	-	-
0.3993	1320	0.8211	-	-
0.4023	1330	0.9117	-	-
0.4053	1340	0.832	-	-
0.4083	1350	0.8222	-	-
0.4114	1360	0.8366	-	-
0.4144	1370	0.871	-	-
0.4174	1380	0.8787	-	-
0.4204	1390	0.8797	-	-
0.4235	1400	0.8222	-	-
0.4265	1410	0.8187	-	-
0.4295	1420	0.9012	-	-
0.4325	1430	0.8047	-	-
0.4356	1440	0.8916	-	-
0.4386	1450	0.9724	-	-
0.4416	1460	0.8306	-	-
0.4446	1470	0.8336	-	-
0.4477	1480	0.8542	-	-
0.4507	1490	0.9075	-	-
0.4537	1500	0.7568	-	-
0.4567	1510	0.9213	-	-
0.4598	1520	0.9079	-	-
0.4628	1530	0.8843	-	-
0.4658	1540	0.8893	-	-
0.4688	1550	0.8085	-	-
0.4719	1560	0.9153	-	-
0.4749	1570	0.851	-	-
0.4779	1580	0.8272	-	-
0.4809	1590	0.8105	-	-
0.4840	1600	0.8512	-	-
0.4870	1610	0.8795	-	-
0.4900	1620	0.7917	-	-
0.4930	1630	0.8111	-	-
0.4961	1640	0.8039	-	-
0.4991	1650	0.8209	-	-
0.5021	1660	0.9045	-	-
0.5051	1670	0.8906	-	-
0.5082	1680	0.8735	-	-
0.5112	1690	0.8643	-	-
0.5142	1700	0.9011	-	-
0.5172	1710	0.9391	-	-
0.5203	1720	0.8082	-	-
0.5233	1730	0.8096	-	-
0.5263	1740	0.883	-	-
0.5293	1750	0.8514	-	-
0.5324	1760	0.8291	-	-
0.5354	1770	0.7463	-	-
0.5384	1780	0.8582	-	-
0.5414	1790	0.9219	-	-
0.5445	1800	0.7607	-	-
0.5475	1810	0.8536	-	-
0.5505	1820	0.7858	-	-
0.5535	1830	0.8204	-	-
0.5566	1840	0.8731	-	-
0.5596	1850	0.8658	-	-
0.5626	1860	0.8901	-	-
0.5656	1870	0.8024	-	-
0.5687	1880	0.8523	-	-
0.5717	1890	0.9049	-	-
0.5747	1900	0.8477	-	-
0.5777	1910	0.7412	-	-
0.5808	1920	0.8318	-	-
0.5838	1930	0.7609	-	-
0.5868	1940	0.7897	-	-
0.5898	1950	0.7879	-	-
0.5929	1960	0.8383	-	-
0.5959	1970	0.8622	-	-
0.5989	1980	0.8009	-	-
0.6019	1990	0.8361	-	-
0.6050	2000	0.8168	-	-
0.6080	2010	0.8514	-	-
0.6110	2020	0.7768	-	-
0.6140	2030	0.8155	-	-
0.6171	2040	0.761	-	-
0.6201	2050	0.8684	-	-
0.6231	2060	0.7832	-	-
0.6261	2070	0.8675	-	-
0.6292	2080	0.8899	-	-
0.6322	2090	0.8539	-	-
0.6352	2100	0.8412	-	-
0.6382	2110	0.8548	-	-
0.6413	2120	0.8051	-	-
0.6443	2130	0.8137	-	-
0.6473	2140	0.8693	-	-
0.6503	2150	0.8512	-	-
0.6534	2160	0.7665	-	-
0.6564	2170	0.7902	-	-
0.6594	2180	0.8232	-	-
0.6624	2190	0.8493	-	-
0.6655	2200	0.8412	-	-
0.6685	2210	0.8504	-	-
0.6715	2220	0.8341	-	-
0.6745	2230	0.797	-	-
0.6776	2240	0.7935	-	-
0.6806	2250	0.8604	-	-
0.6836	2260	0.8726	-	-
0.6866	2270	0.8141	-	-
0.6897	2280	0.8169	-	-
0.6927	2290	0.8585	-	-
0.6957	2300	0.8637	-	-
0.6987	2310	0.8091	-	-
0.7018	2320	0.8252	-	-
0.7048	2330	0.8887	-	-
0.7078	2340	0.7881	-	-
0.7108	2350	0.9142	-	-
0.7139	2360	0.8151	-	-
0.7169	2370	0.8422	-	-
0.7199	2380	0.8303	-	-
0.7229	2390	0.8372	-	-
0.7260	2400	0.8334	-	-
0.7290	2410	0.7451	-	-
0.7320	2420	0.8585	-	-
0.7350	2430	0.8679	-	-
0.7381	2440	0.8344	-	-
0.7411	2450	0.8634	-	-
0.7441	2460	0.7852	-	-
0.7471	2470	0.8399	-	-
0.7502	2480	0.8177	-	-
0.7532	2490	0.9176	-	-
0.7562	2500	0.7577	-	-
0.7592	2510	0.6894	-	-
0.7623	2520	0.8084	-	-
0.7653	2530	0.898	-	-
0.7683	2540	0.8209	-	-
0.7713	2550	0.8621	-	-
0.7743	2560	0.7687	-	-
0.7774	2570	0.8408	-	-
0.7804	2580	0.8467	-	-
0.7834	2590	0.8426	-	-
0.7864	2600	0.8499	-	-
0.7895	2610	0.8973	-	-
0.7925	2620	0.8532	-	-
0.7955	2630	0.8833	-	-
0.7985	2640	0.8137	-	-
0.8016	2650	0.8812	-	-
0.8046	2660	0.8146	-	-
0.8076	2670	0.8285	-	-
0.8106	2680	0.8989	-	-
0.8137	2690	0.8399	-	-
0.8167	2700	0.7851	-	-
0.8197	2710	0.7952	-	-
0.8227	2720	0.7762	-	-
0.8258	2730	0.8184	-	-
0.8288	2740	0.8423	-	-
0.8318	2750	0.8314	-	-
0.8348	2760	0.8078	-	-
0.8379	2770	0.837	-	-
0.8409	2780	0.7494	-	-
0.8439	2790	0.8687	-	-
0.8469	2800	0.8844	-	-
0.8500	2810	0.772	-	-
0.8530	2820	0.8961	-	-
0.8560	2830	0.8599	-	-
0.8590	2840	0.7936	-	-
0.8621	2850	0.8054	-	-
0.8651	2860	0.7812	-	-
0.8681	2870	0.8175	-	-
0.8711	2880	0.8121	-	-
0.8742	2890	0.8192	-	-
0.8772	2900	0.8704	-	-
0.8802	2910	0.8535	-	-
0.8832	2920	0.8187	-	-
0.8863	2930	0.8356	-	-
0.8893	2940	0.835	-	-
0.8923	2950	0.8279	-	-
0.8953	2960	0.8496	-	-
0.8984	2970	0.7985	-	-
0.9014	2980	0.8032	-	-
0.9044	2990	0.8687	-	-
0.9074	3000	0.7948	-	-
0.9105	3010	0.863	-	-
0.9135	3020	0.8589	-	-
0.9165	3030	0.7393	-	-
0.9195	3040	0.7791	-	-
0.9226	3050	0.8215	-	-
0.9256	3060	0.8034	-	-
0.9286	3070	0.8889	-	-
0.9316	3080	0.7151	-	-
0.9347	3090	0.8857	-	-
0.9377	3100	0.8059	-	-
0.9407	3110	0.8435	-	-
0.9437	3120	0.7731	-	-
0.9468	3130	0.8757	-	-
0.9498	3140	0.8846	-	-
0.9528	3150	0.8533	-	-
0.9558	3160	0.8337	-	-
0.9589	3170	0.8618	-	-
0.9619	3180	0.7963	-	-
0.9649	3190	0.7544	-	-
0.9679	3200	0.787	-	-
0.9710	3210	0.7714	-	-
0.9740	3220	0.8506	-	-
0.9770	3230	0.8075	-	-
0.9800	3240	0.9149	-	-
0.9831	3250	0.7584	-	-
0.9861	3260	0.8148	-	-
0.9891	3270	0.7667	-	-
0.9921	3280	0.7781	-	-
0.9952	3290	0.705	-	-
0.9982	3300	0.7874	-	-
1.0	3306	-	0.8084	0.4390

Framework Versions

• Python: 3.12.12
• Sentence Transformers: 5.1.2
• Transformers: 4.57.3
• PyTorch: 2.9.0+cu126
• Accelerate: 1.12.0
• Datasets: 4.0.0
• Tokenizers: 0.22.1

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",
}