afiyarah/gemma-ins-make

SentenceTransformer based on google/embeddinggemma-300m

This is a sentence-transformers model finetuned from google/embeddinggemma-300m. 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: google/embeddinggemma-300m
• Maximum Sequence Length: 2048 tokens
• Output Dimensionality: 768 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': 2048, 'do_lower_case': False, 'architecture': 'Gemma3TextModel'})
  (1): Pooling({'word_embedding_dimension': 768, '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): Dense({'in_features': 768, 'out_features': 3072, 'bias': False, 'activation_function': 'torch.nn.modules.linear.Identity'})
  (3): Dense({'in_features': 3072, 'out_features': 768, 'bias': False, 'activation_function': 'torch.nn.modules.linear.Identity'})
  (4): 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("sentence_transformers_model_id")
# Run inference
queries = [
    "In the car insurance domain, represent this car make entity in arabic and english for entity similarity matching: sym",
]
documents = [
    'In the car insurance domain, represent this car make entity in arabic and english for entity similarity matching: تاتا',
    'In the car insurance domain, represent this car make entity in arabic and english for entity similarity matching: mcv',
    'In the car insurance domain, represent this car make entity in arabic for entity similarity matching: تورو',
]
query_embeddings = model.encode_query(queries)
document_embeddings = model.encode_document(documents)
print(query_embeddings.shape, document_embeddings.shape)
# [1, 768] [3, 768]

# Get the similarity scores for the embeddings
similarities = model.similarity(query_embeddings, document_embeddings)
print(similarities)
# tensor([[0.4455, 0.4351, 0.5173]])

Evaluation

Metrics

Semantic Similarity

• Dataset: insurance-val
• Evaluated with EmbeddingSimilarityEvaluator

Metric	Value
pearson_cosine	0.9518
spearman_cosine	0.9267

Training Details

Training Dataset

Unnamed Dataset

• Size: 33,973 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: 21 tokens mean: 23.89 tokens max: 29 tokens	min: 21 tokens mean: 23.88 tokens max: 29 tokens	min: 0.0 mean: 0.2 max: 1.0

• Samples:

  sentence_0	sentence_1	label
  In the car insurance domain, represent this car make entity in arabic and english for entity similarity matching: كوبرا	In the car insurance domain, represent this car make entity in arabic and english for entity similarity matching: cobra	1.0
  In the car insurance domain, represent this car make entity in arabic and english for entity similarity matching: رووتس ملتي كلين	In the car insurance domain, represent this car make entity in arabic and english for entity similarity matching: rokon	0.0
  In the car insurance domain, represent this car make entity in arabic and english for entity similarity matching: شاينراي	In the car insurance domain, represent this car make entity in arabic and english for entity similarity matching: shineray	1.0

• 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: 5
• 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: 1.0
• num_train_epochs: 5
• 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: 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
• 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: False
• 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: False
• prompts: None
• batch_sampler: batch_sampler
• multi_dataset_batch_sampler: round_robin
• router_mapping: {}
• learning_rate_mapping: {}

Training Logs

Epoch	Step	Training Loss	insurance-val_spearman_cosine
0.1996	424	-	0.6773
0.2354	500	0.0477	-
0.3992	848	-	0.7524
0.4708	1000	0.0237	-
0.5989	1272	-	0.7464
0.7062	1500	0.0234	-
0.7985	1696	-	0.8396
0.9416	2000	0.0182	-
0.9981	2120	-	0.8616
1.0	2124	-	0.8494
1.1770	2500	0.0122	-
1.1977	2544	-	0.8766
1.3974	2968	-	0.8274
1.4124	3000	0.0107	-
1.5970	3392	-	0.8692
1.6478	3500	0.0087	-
1.7966	3816	-	0.8744
1.8832	4000	0.0085	-
1.9962	4240	-	0.8921
2.0	4248	-	0.8943
2.1186	4500	0.0059	-
2.1959	4664	-	0.8954
2.3540	5000	0.0043	-
2.3955	5088	-	0.9033
2.5895	5500	0.0038	-
2.5951	5512	-	0.8952
2.7947	5936	-	0.8996
2.8249	6000	0.0035	-
2.9944	6360	-	0.9217
3.0	6372	-	0.9214
3.0603	6500	0.0028	-
3.1940	6784	-	0.9259
3.2957	7000	0.0018	-
3.3936	7208	-	0.9242
3.5311	7500	0.0019	-
3.5932	7632	-	0.9186
3.7665	8000	0.0019	-
3.7928	8056	-	0.9172
3.9925	8480	-	0.9195
4.0	8496	-	0.9193
4.0019	8500	0.0016	-
4.1921	8904	-	0.9203
4.2373	9000	0.001	-
4.3917	9328	-	0.9248
4.4727	9500	0.0008	-
4.5913	9752	-	0.9264
4.7081	10000	0.0009	-
4.7910	10176	-	0.9264
4.9435	10500	0.001	-
4.9906	10600	-	0.9267

Framework Versions

• Python: 3.12.11
• Sentence Transformers: 5.1.0
• Transformers: 4.56.1
• PyTorch: 2.8.0+cu126
• Accelerate: 1.10.1
• Datasets: 4.0.0
• Tokenizers: 0.22.0

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",
}