JatinkInnovision/snowflake-arctic-embed-l-v2.0_all-nli

SentenceTransformer based on Snowflake/snowflake-arctic-embed-l-v2.0

This is a sentence-transformers model finetuned from Snowflake/snowflake-arctic-embed-l-v2.0 on the all-nli dataset. 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: Snowflake/snowflake-arctic-embed-l-v2.0
• Maximum Sequence Length: 8192 tokens
• Output Dimensionality: 1024 dimensions
• Similarity Function: Cosine Similarity
• Training Dataset:
  • all-nli
• Language: en

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}) with Transformer model: XLMRobertaModel 
  (1): Pooling({'word_embedding_dimension': 1024, '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})
  (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("JatinkInnovision/snowflake-arctic-embed-l-v2.0_all-nli")
# Run inference
sentences = [
    'A middle-aged man works under the engine of a train on rail tracks.',
    'A guy is working on a train.',
    'A guy is driving to work.',
]
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

Triplet

• Dataset: all-nli-test
• Evaluated with TripletEvaluator

Metric	Value
cosine_accuracy	0.9558

Training Details

Training Dataset

all-nli

• Dataset: all-nli at d482672
• 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: 7 tokens mean: 10.9 tokens max: 52 tokens	min: 6 tokens mean: 13.62 tokens max: 42 tokens	min: 5 tokens mean: 14.76 tokens max: 55 tokens

• Samples:

  anchor	positive	negative
  A person on a horse jumps over a broken down airplane.	A person is outdoors, on a horse.	A person is at a diner, ordering an omelette.
  Children smiling and waving at camera	There are children present	The kids are frowning
  A boy is jumping on skateboard in the middle of a red bridge.	The boy does a skateboarding trick.	The boy skates down the sidewalk.

• Loss: MultipleNegativesRankingLoss with these parameters:

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

Evaluation Dataset

all-nli

• Dataset: all-nli at d482672
• 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: 6 tokens mean: 20.31 tokens max: 83 tokens	min: 5 tokens mean: 10.71 tokens max: 35 tokens	min: 5 tokens mean: 11.39 tokens max: 32 tokens

• Samples:

  anchor	positive	negative
  Two women are embracing while holding to go packages.	Two woman are holding packages.	The men are fighting outside a deli.
  Two young children in blue jerseys, one with the number 9 and one with the number 2 are standing on wooden steps in a bathroom and washing their hands in a sink.	Two kids in numbered jerseys wash their hands.	Two kids in jackets walk to school.
  A man selling donuts to a customer during a world exhibition event held in the city of Angeles	A man selling donuts to a customer.	A woman drinks her coffee in a small cafe.

• Loss: MultipleNegativesRankingLoss with these parameters:

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

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 50
• per_device_eval_batch_size: 50
• num_train_epochs: 1
• warmup_ratio: 0.1
• fp16: 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: 50
• per_device_eval_batch_size: 50
• 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: 1
• 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
• 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: None
• 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

Click to expand

Epoch	Step	Training Loss	Validation Loss	all-nli-test_cosine_accuracy
0.0090	100	1.8838	0.6502	-
0.0179	200	1.2991	0.6177	-
0.0269	300	1.2721	0.6417	-
0.0359	400	1.2265	0.7053	-
0.0448	500	1.0111	0.7147	-
0.0538	600	1.0491	0.7457	-
0.0627	700	1.0186	0.7922	-
0.0717	800	1.135	0.8940	-
0.0807	900	1.0747	0.7007	-
0.0896	1000	0.9373	0.7298	-
0.0986	1100	0.9572	0.6809	-
0.1076	1200	1.1316	0.7260	-
0.1165	1300	0.9188	0.7085	-
0.1255	1400	0.9554	0.6876	-
0.1344	1500	0.9494	0.7492	-
0.1434	1600	0.811	0.7234	-
0.1524	1700	0.7766	0.6744	-
0.1613	1800	0.9317	0.7178	-
0.1703	1900	0.9148	0.6960	-
0.1793	2000	0.8643	0.6642	-
0.1882	2100	0.7604	0.6425	-
0.1972	2200	0.776	0.6347	-
0.2061	2300	0.8286	0.6581	-
0.2151	2400	0.8946	0.5866	-
0.2241	2500	0.8507	0.6845	-
0.2330	2600	0.7917	0.6091	-
0.2420	2700	0.8192	0.7073	-
0.2510	2800	0.8818	0.6584	-
0.2599	2900	0.8261	0.6112	-
0.2689	3000	0.8017	0.6883	-
0.2779	3100	0.8147	0.6450	-
0.2868	3200	0.8297	0.6086	-
0.2958	3300	0.7516	0.5857	-
0.3047	3400	0.8628	0.6061	-
0.3137	3500	0.7758	0.5751	-
0.3227	3600	0.7773	0.6022	-
0.3316	3700	0.7559	0.5446	-
0.3406	3800	0.796	0.5842	-
0.3496	3900	0.8295	0.5822	-
0.3585	4000	0.7292	0.5821	-
0.3675	4100	0.7475	0.6358	-
0.3764	4200	0.7916	0.5688	-
0.3854	4300	0.7214	0.5653	-
0.3944	4400	0.704	0.5564	-
0.4033	4500	0.7817	0.5876	-
0.4123	4600	0.7549	0.5358	-
0.4213	4700	0.7206	0.5785	-
0.4302	4800	0.7462	0.5568	-
0.4392	4900	0.665	0.5765	-
0.4481	5000	0.7743	0.5303	-
0.4571	5100	0.7055	0.5733	-
0.4661	5200	0.7004	0.6280	-
0.4750	5300	0.7021	0.5444	-
0.4840	5400	0.6858	0.5787	-
0.4930	5500	0.7007	0.6124	-
0.5019	5600	0.6722	0.5705	-
0.5109	5700	0.7124	0.5440	-
0.5199	5800	0.6657	0.5262	-
0.5288	5900	0.6784	0.5400	-
0.5378	6000	0.6644	0.5093	-
0.5467	6100	0.7195	0.5453	-
0.5557	6200	0.6958	0.5216	-
0.5647	6300	0.7202	0.5250	-
0.5736	6400	0.6921	0.5089	-
0.5826	6500	0.6926	0.5207	-
0.5916	6600	0.714	0.5084	-
0.6005	6700	0.6605	0.4943	-
0.6095	6800	0.7222	0.5058	-
0.6184	6900	0.7171	0.4950	-
0.6274	7000	0.6344	0.5110	-
0.6364	7100	0.7057	0.5197	-
0.6453	7200	0.6895	0.5096	-
0.6543	7300	0.7226	0.4819	-
0.6633	7400	0.6725	0.4780	-
0.6722	7500	0.7469	0.5145	-
0.6812	7600	0.7016	0.4969	-
0.6901	7700	0.6655	0.4965	-
0.6991	7800	0.7281	0.4913	-
0.7081	7900	0.6748	0.5121	-
0.7170	8000	0.6505	0.5207	-
0.7260	8100	0.6594	0.4823	-
0.7350	8200	0.7042	0.4903	-
0.7439	8300	0.6995	0.4630	-
0.7529	8400	0.634	0.4217	-
0.7619	8500	0.3772	0.3684	-
0.7708	8600	0.3416	0.3585	-
0.7798	8700	0.3113	0.3471	-
0.7887	8800	0.2793	0.3379	-
0.7977	8900	0.2577	0.3349	-
0.8067	9000	0.249	0.3320	-
0.8156	9100	0.2191	0.3290	-
0.8246	9200	0.2492	0.3255	-
0.8336	9300	0.2464	0.3258	-
0.8425	9400	0.2288	0.3247	-
0.8515	9500	0.2132	0.3248	-
0.8604	9600	0.2173	0.3259	-
0.8694	9700	0.2008	0.3223	-
0.8784	9800	0.2016	0.3219	-
0.8873	9900	0.1962	0.3195	-
0.8963	10000	0.1952	0.3185	-
0.9053	10100	0.1959	0.3158	-
0.9142	10200	0.2002	0.3138	-
0.9232	10300	0.1882	0.3150	-
0.9322	10400	0.1856	0.3124	-
0.9411	10500	0.1971	0.3143	-
0.9501	10600	0.1918	0.3137	-
0.9590	10700	0.1825	0.3147	-
0.9680	10800	0.1762	0.3155	-
0.9770	10900	0.1778	0.3139	-
0.9859	11000	0.1659	0.3138	-
0.9949	11100	0.1848	0.3131	-
1.0	11157	-	-	0.9558

Framework Versions

• Python: 3.10.12
• Sentence Transformers: 3.3.1
• Transformers: 4.47.1
• PyTorch: 2.5.1+cu121
• Accelerate: 1.2.1
• Datasets: 3.2.0
• Tokenizers: 0.21.0

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",
}

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}
}