SentenceTransformer based on BAAI/bge-base-en-v1.5
This is a sentence-transformers model finetuned from BAAI/bge-base-en-v1.5. 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: BAAI/bge-base-en-v1.5
- Maximum Sequence Length: 512 tokens
- Output Dimensionality: 768 dimensions
- Similarity Function: Cosine Similarity
Model Sources
Full Model Architecture
SentenceTransformer(
(0): Transformer({'max_seq_length': 512, 'do_lower_case': True}) with Transformer model: BertModel
(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})
(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
model = SentenceTransformer("ayushexel/emb-bge-base-en-v1.5-squad-3-epochs")
sentences = [
'How many people outside the UK were under British rule in 1945?',
'Though Britain and the empire emerged victorious from the Second World War, the effects of the conflict were profound, both at home and abroad. Much of Europe, a continent that had dominated the world for several centuries, was in ruins, and host to the armies of the United States and the Soviet Union, who now held the balance of global power. Britain was left essentially bankrupt, with insolvency only averted in 1946 after the negotiation of a $US 4.33 billion loan (US$56 billion in 2012) from the United States, the last instalment of which was repaid in 2006. At the same time, anti-colonial movements were on the rise in the colonies of European nations. The situation was complicated further by the increasing Cold War rivalry of the United States and the Soviet Union. In principle, both nations were opposed to European colonialism. In practice, however, American anti-communism prevailed over anti-imperialism, and therefore the United States supported the continued existence of the British Empire to keep Communist expansion in check. The "wind of change" ultimately meant that the British Empire\'s days were numbered, and on the whole, Britain adopted a policy of peaceful disengagement from its colonies once stable, non-Communist governments were available to transfer power to. This was in contrast to other European powers such as France and Portugal, which waged costly and ultimately unsuccessful wars to keep their empires intact. Between 1945 and 1965, the number of people under British rule outside the UK itself fell from 700 million to five million, three million of whom were in Hong Kong.',
"By the start of the 20th century, Germany and the United States challenged Britain's economic lead. Subsequent military and economic tensions between Britain and Germany were major causes of the First World War, during which Britain relied heavily upon its empire. The conflict placed enormous strain on the military, financial and manpower resources of Britain. Although the British Empire achieved its largest territorial extent immediately after World War I, Britain was no longer the world's pre-eminent industrial or military power. In the Second World War, Britain's colonies in South-East Asia were occupied by Imperial Japan. Despite the final victory of Britain and its allies, the damage to British prestige helped to accelerate the decline of the empire. India, Britain's most valuable and populous possession, achieved independence as part of a larger decolonisation movement in which Britain granted independence to most territories of the Empire. The transfer of Hong Kong to China in 1997 marked for many the end of the British Empire. Fourteen overseas territories remain under British sovereignty. After independence, many former British colonies joined the Commonwealth of Nations, a free association of independent states. The United Kingdom is now one of 16 Commonwealth nations, a grouping known informally as the Commonwealth realms, that share one monarch—Queen Elizabeth II.",
]
embeddings = model.encode(sentences)
print(embeddings.shape)
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
Evaluation
Metrics
Triplet
| Metric |
Value |
| cosine_accuracy |
0.4192 |
Training Details
Training Dataset
Unnamed Dataset
Evaluation Dataset
Unnamed Dataset
Training Hyperparameters
Non-Default Hyperparameters
eval_strategy: steps
per_device_train_batch_size: 128
per_device_eval_batch_size: 128
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: 128
per_device_eval_batch_size: 128
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: 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
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}
tp_size: 0
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
| Epoch |
Step |
Training Loss |
Validation Loss |
gooqa-dev_cosine_accuracy |
| -1 |
-1 |
- |
- |
0.3536 |
| 0.2890 |
100 |
0.6663 |
0.7830 |
0.3922 |
| 0.5780 |
200 |
0.4374 |
0.7399 |
0.4064 |
| 0.8671 |
300 |
0.4048 |
0.7294 |
0.4086 |
| 1.1561 |
400 |
0.3149 |
0.7244 |
0.4136 |
| 1.4451 |
500 |
0.2378 |
0.7246 |
0.4182 |
| 1.7341 |
600 |
0.2358 |
0.7179 |
0.4158 |
| 2.0231 |
700 |
0.2338 |
0.7170 |
0.4240 |
| 2.3121 |
800 |
0.1602 |
0.7293 |
0.4148 |
| 2.6012 |
900 |
0.1595 |
0.7237 |
0.4230 |
| 2.8902 |
1000 |
0.1545 |
0.7229 |
0.4146 |
| -1 |
-1 |
- |
- |
0.4192 |
Framework Versions
- Python: 3.11.0
- Sentence Transformers: 4.0.1
- Transformers: 4.50.3
- PyTorch: 2.6.0+cu124
- Accelerate: 1.5.2
- Datasets: 3.5.0
- Tokenizers: 0.21.1
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",
}
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}
}