This is a sentence-transformers model finetuned from kiarashmo/chembberta-77m-mlm-safetensors. It maps sentences & paragraphs to a 384-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: kiarashmo/chembberta-77m-mlm-safetensors
Maximum Sequence Length: 512 tokens
Output Dimensionality: 384 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': 512, 'do_lower_case': False, 'architecture': 'RobertaModel'})
  (1): Pooling({'word_embedding_dimension': 384, '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})
)

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.

python

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("sentence_transformers_model_id")
# Run inference
sentences = [
    'COCC(=O)NC(Cc1cccc(-c2nccn2C)c1)C(O)C[NH2+]C1CC2(CCC2)Oc2ncc(CC(C)(C)C)cc21',
    'C#CC[NH2+]CC(O)C(Cc1ccccc1)NC(=O)c1cc2c3c(c1)c(CC)cn3CCS(=O)(=O)O2',
    'CC1(c2cc(NC(=O)c3ncc(Cl)cn3)ccc2F)N=C(N)OCC1(F)F',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 384]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities)
# tensor([[1.0000, 0.8032, 0.8972],
#         [0.8032, 1.0000, 0.8267],
#         [0.8972, 0.8267, 1.0000]])

Evaluation

Metrics

Binary Classification

Dataset: val-sim
Evaluated with BinaryClassificationEvaluator

Metric	Value
cosine_accuracy	0.574
cosine_accuracy_threshold	0.3659
cosine_f1	0.6699
cosine_f1_threshold	0.0002
cosine_precision	0.5105
cosine_recall	0.9741
cosine_ap	0.5801
cosine_mcc	0.0882

Training Details

Training Dataset

Unnamed Dataset

Size: 30,000 training samples
Columns: sentence_A, sentence_B, and label
Approximate statistics based on the first 1000 samples:
sentence_A sentence_B label
type string string int
details
min: 17 tokens
mean: 61.05 tokens
max: 129 tokens
min: 17 tokens
mean: 61.82 tokens
max: 123 tokens
0: ~52.90%
1: ~47.10%

Samples:

sentence_A	sentence_B	label
`CN1C(=O)C(c2ccc(OC(F)F)cc2)(c2cccc(C#CCCCCO)c2)N=C1N`	`COc1cccc(C[NH2+]CC(O)C(Cc2cc(F)cc(F)c2)NC(=O)c2cc(C(C)=O)cc(N(C)S(C)(=O)=O)c2)c1`	`0`
`CN1C(=O)C(c2ccccc2)(c2cccc(Cl)c2)N=C1N`	`CCCOC1C[NH2+]C(C(O)C(Cc2cc(F)cc(F)c2)NC(=O)C(C)N2CCC(C(C)C)C2=O)C1`	`1`
`CCC(CC(=O)NC(COC(C)(C)C)C(=O)[O-])n1c(N)nc2cc(Cl)ccc21`	`CN1C(=O)C(c2ccccc2)(C23CC4CC(CC(C4)C2)C3)N=C1N`	`1`

Loss: SoftmaxLoss

Evaluation Dataset

Unnamed Dataset

Size: 1,000 evaluation samples
Columns: sentence_A, sentence_B, and label
Approximate statistics based on the first 1000 samples:
sentence_A sentence_B label
type string string int
details
min: 46 tokens
mean: 63.21 tokens
max: 110 tokens
min: 46 tokens
mean: 63.68 tokens
max: 110 tokens
0: ~49.90%
1: ~50.10%

Samples:

sentence_A	sentence_B	label
`CCNc1cc(C(=O)NC(Cc2ccccc2)C(O)C[NH2+]C2CC2)cc(N2CCCCS2(=O)=O)c1`	`CC(=O)NC(Cc1cc(F)cc(F)c1)C(O)C[NH2+]C1(c2cccc(C3CCOC3)c2)CCOCC1`	`1`
`CC(=O)NC(Cc1cc(F)cc(F)c1)C(O)C[NH2+]C1(c2cccc(C(C)(C)C)c2)CCC(Nc2cccnc2)CC1`	`CN1C(=O)C(c2ccsc2)(c2cccc(-c3cnccn3)c2)N=C1N`	`0`
`CC(=O)NC(Cc1cc(F)cc(F)c1)C(O)C[NH2+]C1(c2cccc(C(C)(C)C)c2)CCC(=O)CC1`	`CC#Cc1ccnc(-c2cc(C3(C4CC4)N=C(C)C(N)=N3)ccc2O)c1`	`1`

Loss: SoftmaxLoss

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: steps
per_device_train_batch_size: 32
per_device_eval_batch_size: 32
num_train_epochs: 100
warmup_steps: 100
load_best_model_at_end: True

All Hyperparameters

Click to expand

overwrite_output_dir: False
do_predict: False
eval_strategy: steps
prediction_loss_only: True
per_device_train_batch_size: 32
per_device_eval_batch_size: 32
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: 100
max_steps: -1
lr_scheduler_type: linear
lr_scheduler_kwargs: {}
warmup_ratio: 0.0
warmup_steps: 100
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: 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: proportional
router_mapping: {}
learning_rate_mapping: {}

Training Logs

Epoch	Step	Training Loss	Validation Loss	val-sim_cosine_ap
0.533	500	0.6775	0.7206	0.5711
1.0661	1000	0.4513	0.9449	0.5901
1.5991	1500	0.3204	1.1887	0.5995
2.1322	2000	0.247	1.2793	0.5886
2.6652	2500	0.201	1.4287	0.5969
3.1983	3000	0.1703	1.5783	0.5801

The bold row denotes the saved checkpoint.

Framework Versions

Python: 3.9.23
Sentence Transformers: 5.0.0
Transformers: 4.53.3
PyTorch: 2.5.0+cu118
Accelerate: 1.9.0
Datasets: 4.0.0
Tokenizers: 0.21.2

Citation

BibTeX

Sentence Transformers and SoftmaxLoss

bibtex

@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",
}

	sentence_A	sentence_B	label
type	string	string	int
details	min: 17 tokens mean: 61.05 tokens max: 129 tokens	min: 17 tokens mean: 61.82 tokens max: 123 tokens	0: ~52.90% 1: ~47.10%

kiarashmo/chembberta-77m-sBERT-finetuned-on-bace-with-sBERTSoftmaxLoss

SentenceTransformer based on kiarashmo/chembberta-77m-mlm-safetensors

Model Details

Model Description

Model Sources

Full Model Architecture

Usage

Direct Usage (Sentence Transformers)

Evaluation

Metrics

Binary Classification

Training Details

Training Dataset

Unnamed Dataset

Evaluation Dataset

Unnamed Dataset

Training Hyperparameters

Non-Default Hyperparameters

All Hyperparameters

Training Logs

Framework Versions

Citation

BibTeX

Sentence Transformers and SoftmaxLoss

kiarashmo/
chembberta-77m-sBERT-finetuned-on-bace-with-sBERTSoftmaxLoss