SentenceTransformer based on jinaai/jina-embeddings-v3

This is a sentence-transformers model finetuned from jinaai/jina-embeddings-v3. It maps sentences & paragraphs to a 1024-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: jinaai/jina-embeddings-v3
Maximum Sequence Length: 8194 tokens
Output Dimensionality: 1024 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(
  (transformer): Transformer(
    (auto_model): XLMRobertaLoRA(
      (roberta): XLMRobertaModel(
        (embeddings): XLMRobertaEmbeddings(
          (word_embeddings): ParametrizedEmbedding(
            250002, 1024, padding_idx=1
            (parametrizations): ModuleDict(
              (weight): ParametrizationList(
                (0): LoRAParametrization()
              )
            )
          )
          (token_type_embeddings): ParametrizedEmbedding(
            1, 1024
            (parametrizations): ModuleDict(
              (weight): ParametrizationList(
                (0): LoRAParametrization()
              )
            )
          )
        )
        (emb_drop): Dropout(p=0.1, inplace=False)
        (emb_ln): LayerNorm((1024,), eps=1e-05, elementwise_affine=True)
        (encoder): XLMRobertaEncoder(
          (layers): ModuleList(
            (0-23): 24 x Block(
              (mixer): MHA(
                (rotary_emb): RotaryEmbedding()
                (Wqkv): ParametrizedLinearResidual(
                  in_features=1024, out_features=3072, bias=True
                  (parametrizations): ModuleDict(
                    (weight): ParametrizationList(
                      (0): LoRAParametrization()
                    )
                  )
                )
                (inner_attn): SelfAttention(
                  (drop): Dropout(p=0.1, inplace=False)
                )
                (inner_cross_attn): CrossAttention(
                  (drop): Dropout(p=0.1, inplace=False)
                )
                (out_proj): ParametrizedLinear(
                  in_features=1024, out_features=1024, bias=True
                  (parametrizations): ModuleDict(
                    (weight): ParametrizationList(
                      (0): LoRAParametrization()
                    )
                  )
                )
              )
              (dropout1): Dropout(p=0.1, inplace=False)
              (drop_path1): StochasticDepth(p=0.0, mode=row)
              (norm1): LayerNorm((1024,), eps=1e-05, elementwise_affine=True)
              (mlp): Mlp(
                (fc1): ParametrizedLinear(
                  in_features=1024, out_features=4096, bias=True
                  (parametrizations): ModuleDict(
                    (weight): ParametrizationList(
                      (0): LoRAParametrization()
                    )
                  )
                )
                (fc2): ParametrizedLinear(
                  in_features=4096, out_features=1024, bias=True
                  (parametrizations): ModuleDict(
                    (weight): ParametrizationList(
                      (0): LoRAParametrization()
                    )
                  )
                )
              )
              (dropout2): Dropout(p=0.1, inplace=False)
              (drop_path2): StochasticDepth(p=0.0, mode=row)
              (norm2): LayerNorm((1024,), eps=1e-05, elementwise_affine=True)
            )
          )
        )
        (pooler): XLMRobertaPooler(
          (dense): ParametrizedLinear(
            in_features=1024, out_features=1024, bias=True
            (parametrizations): ModuleDict(
              (weight): ParametrizationList(
                (0): LoRAParametrization()
              )
            )
          )
          (activation): Tanh()
        )
      )
    )
  )
  (pooler): Pooling({'word_embedding_dimension': 1024, '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})
  (normalizer): 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("seregadgl/t2")
# Run inference
sentences = [
    'honor watch gs pro black ',
    'honor watch gs pro white ',
    'трансформер pituso carlo hb gy 06 lemon',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 1024]

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

Evaluation

Metrics

Semantic Similarity

Dataset: example-dev
Evaluated with EmbeddingSimilarityEvaluator

Metric	Value
pearson_cosine	0.4774
spearman_cosine	0.4969

Training Details

Training Dataset

Unnamed Dataset

Size: 63,802 training samples
Columns: doc, candidate, and label
Approximate statistics based on the first 1000 samples:
doc candidate label
type string string int
details
min: 3 tokens
mean: 14.82 tokens
max: 55 tokens

min: 4 tokens
mean: 14.58 tokens
max: 37 tokens

0: ~85.20%
1: ~14.80%

	doc	candidate	label
type	string	string	int
details	min: 3 tokens mean: 14.82 tokens max: 55 tokens	min: 4 tokens mean: 14.58 tokens max: 37 tokens	0: ~85.20% 1: ~14.80%

Samples:

doc	candidate	label
`массажер xiaomi massage gun eu bhr5608eu`	`перкуссионный массажер xiaomi massage gun mini bhr6083gl`	`0`
`безударная дрель ingco ed50028`	`ударная дрель ingco id211002`	`0`
`жидкость old smuggler 30мл 20мг`	`жидкость old smuggler salt 30ml marlboro 20mg`	`0`

Loss: CoSENTLoss with these parameters:

{
    "scale": 20.0,
    "similarity_fct": "pairwise_cos_sim"
}

Evaluation Dataset

Unnamed Dataset

Size: 7,090 evaluation samples
Columns: doc, candidate, and label
Approximate statistics based on the first 1000 samples:
doc candidate label
type string string int
details
min: 4 tokens
mean: 14.91 tokens
max: 72 tokens

min: 4 tokens
mean: 14.56 tokens
max: 51 tokens

0: ~84.20%
1: ~15.80%

	doc	candidate	label
type	string	string	int
details	min: 4 tokens mean: 14.91 tokens max: 72 tokens	min: 4 tokens mean: 14.56 tokens max: 51 tokens	0: ~84.20% 1: ~15.80%

Samples:

doc	candidate	label
`круглое пляжное парео селфи коврик пляжная подстилка пляжное покрывало пляжный коврик пироженко`	`круглое пляжное парео селфи коврик пляжная подстилка пляжное покрывало пляжный коврик клубника`	`0`
`аккумулятор батарея для ноутбука asus g751`	`аккумулятор батарея для ноутбука asus g75 series`	`0`
`миксер bosch mfq3520 mfq 3520`	`миксер bosch mfq 4020`	`0`

Loss: CoSENTLoss with these parameters:

{
    "scale": 20.0,
    "similarity_fct": "pairwise_cos_sim"
}

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: steps
per_device_train_batch_size: 16
per_device_eval_batch_size: 16
num_train_epochs: 2
lr_scheduler_type: cosine
warmup_ratio: 0.1
load_best_model_at_end: True
batch_sampler: no_duplicates

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: 2
max_steps: -1
lr_scheduler_type: cosine
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: 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}
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
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
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
use_liger_kernel: False
eval_use_gather_object: False
average_tokens_across_devices: False
prompts: None
batch_sampler: no_duplicates
multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	Validation Loss	example-dev_spearman_cosine
0	0	-	-	0.1562
0.1254	500	4.2363	3.5101	0.3313
0.2508	1000	3.0049	2.8592	0.4536
0.3761	1500	2.6306	2.8977	0.4704
0.5015	2000	2.6472	2.6703	0.4827
0.6269	2500	2.6626	2.6757	0.4837
0.7523	3000	2.6137	2.6397	0.4883
0.8776	3500	2.676	2.5394	0.4936
1.0030	4000	2.4997	2.5984	0.4931
1.1284	4500	2.4901	2.6219	0.4946
1.2538	5000	2.4293	2.6319	0.4943
1.3791	5500	2.3914	2.7122	0.4936
1.5045	6000	2.465	2.6573	0.4970
1.6299	6500	2.5711	2.6388	0.4965
1.7553	7000	2.5012	2.6323	0.4967
1.8806	7500	2.5775	2.6231	0.4969

Framework Versions

Python: 3.10.14
Sentence Transformers: 3.3.1
Transformers: 4.46.3
PyTorch: 2.4.0
Accelerate: 0.34.2
Datasets: 3.0.1
Tokenizers: 0.20.0

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",
}

CoSENTLoss

@online{kexuefm-8847,
    title={CoSENT: A more efficient sentence vector scheme than Sentence-BERT},
    author={Su Jianlin},
    year={2022},
    month={Jan},
    url={https://kexue.fm/archives/8847},
}

Downloads last month: 3

Safetensors

Model size

0.6B params

Tensor type

F32

Model tree for seregadgl/t2

Base model

jinaai/jina-embeddings-v3

Finetuned

(30)

this model

Paper for seregadgl/t2

Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks

Paper • 1908.10084 • Published Aug 27, 2019 • 9

Evaluation results

Pearson Cosine on example dev
self-reported

0.477
Spearman Cosine on example dev
self-reported

0.497