SentenceTransformer based on BAAI/bge-base-en-v1.5

This is a sentence-transformers model finetuned from BAAI/bge-base-en-v1.5 on the csv dataset. 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 tokens
Similarity Function: Cosine Similarity
Training Dataset:
- csv

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': 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

# Download from the 🤗 Hub
model = SentenceTransformer("Gurveer05/bge-base-eedi-2024")
# Run inference
sentences = [
    'Question: Solve quadratic equations using the quadratic formula where the coefficient of x² is 1. Vera wants to solve this equation using the quadratic formula.\n(\nh^2+4=5 h\n)\n\nWhat should replace the triangle? The image shows the structure of the quadratic formula. It says plus or minus the square root, and the triangle is the first thing after the square root sign, with a minus sign after it.\n\nOptions:\nA. 8\nB. -10\nC. 16\nD. 25\n\nAnswer: -10',
    'Mixes up squaring and multiplying by 2 or doubling',
    'Does not realise a quadratic must be in the form ax^2+bx+c=0 to find the values for the quadratic formula',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]

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

Training Details

Training Dataset

csv

Dataset: csv
Size: 2,940 training samples
Columns: anchor, positive, and negative

Approximate statistics based on the first 1000 samples:

	anchor	positive	negative
type	string	string	string
details	min: 33 tokens mean: 90.0 tokens max: 512 tokens	min: 4 tokens mean: 14.71 tokens max: 38 tokens	min: 7 tokens mean: 16.48 tokens max: 39 tokens

Samples:

anchor	positive	negative
`Question: Add algebraic fractions where the denominators are single terms and are not multiples of each other. Express the following as a single fraction, writing your answer as simply as possible: (t / s)+(2 s / t). Options: A. (t^2+4 s^2 / s t) B. (t+2 s / s+t) C. (2 s t / s+t) D. (t^2+2 s^2 / s t) Answer: (2 s t / s+t)`	`When adding/subtracting fractions, adds/subtracts the denominators and multiplies the numerators`	`Thinks can combine the numerator and denominator after simplifying an algebraic fraction`
Question: Calculate the volume of a cone where the dimensions are all given in the same units. STEP 2 Jessica is trying to work out the volume of this cone. A cone with the slant height labelled 9cm, the perpendicular height labelled h and half the cone's base (forming a right angled triangle with the slant and perpendicular heights) is labelled 6cm. First she needs the perpendicular height. Which of the following equations is true? Options: A. h^2=9^2+6^2 B. h^2=9^2-6^2 C. h^2=12^2+9^2 D. h^2=12^2-9^2 Answer: h^2=12^2-9^2	`When using Pythagoras to find the height of an isosceles triangle, uses the whole base instead of half`	`Has multiplied base by slant height and perpendicular height to find area of a triangle`
`Question: Convert from hours to minutes. 3 hours is the same as ___________ minutes. Options: A. 180 B. 90 C. 30 D. 300 Answer: 90`	`Thinks there are 30 minutes in a hour`	`Answers as if there are 100 minutes in an hour when changing from hours to minutes`

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: 24
per_device_eval_batch_size: 24
learning_rate: 2e-05
weight_decay: 0.01
num_train_epochs: 20
lr_scheduler_type: cosine_with_restarts
warmup_ratio: 0.1
fp16: True
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: 24
per_device_eval_batch_size: 24
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: 0.01
adam_beta1: 0.9
adam_beta2: 0.999
adam_epsilon: 1e-08
max_grad_norm: 1.0
num_train_epochs: 20
max_steps: -1
lr_scheduler_type: cosine_with_restarts
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: 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
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
eval_use_gather_object: False
batch_sampler: no_duplicates
multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss
0.2581	16	3.3202
0.5	31	-
0.5161	32	2.9432
0.7742	48	2.6014
1.0	62	-
1.0323	64	2.1029
1.1613	80	1.5757
1.3710	93	-
1.4194	96	2.0139
1.6774	112	1.8208
1.8710	124	-
1.9355	128	1.6599
2.0645	144	0.7017
2.2419	155	-
2.3226	160	1.4833
2.5806	176	1.3274
2.7419	186	-
2.8387	192	1.1951
3.0968	208	0.5799
3.1129	217	-
3.2258	224	0.9517
3.4839	240	1.0177
3.6129	248	-
3.7419	256	0.8864
4.0	272	0.7591
4.1129	279	-
4.1290	288	0.4319
4.3871	304	0.7878
4.4839	310	-
4.6452	320	0.7483
4.9032	336	0.6432
4.9839	341	-
5.0323	352	0.2496
5.2903	368	0.6689
5.3548	372	-
5.5484	384	0.628
5.8065	400	0.4981
5.8548	403	-
6.0645	416	0.3208
6.1935	432	0.4169
6.2258	434	-
6.4516	448	0.5049
6.7097	464	0.4402
6.7258	465	-
6.9677	480	0.3819
7.0968	496	0.1854
7.3548	512	0.4292
7.5968	527	-
7.6129	528	0.4171
7.8710	544	0.318
8.0968	558	-
8.1290	560	0.1318
8.2581	576	0.3829
8.4677	589	-
8.5161	592	0.4097
8.7742	608	0.2676
8.9677	620	-

The bold row denotes the saved checkpoint.

Framework Versions

Python: 3.10.14
Sentence Transformers: 3.1.1
Transformers: 4.44.0
PyTorch: 2.4.0
Accelerate: 0.33.0
Datasets: 2.21.0
Tokenizers: 0.19.1

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}
}