metadata

base_model: microsoft/mpnet-base
datasets: []
language: []
library_name: sentence-transformers
pipeline_tag: sentence-similarity
tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:24901
  - loss:SoftmaxLoss
widget:
  - source_sentence: >-
      Cryptocurrency holders are being exploited, with whales creating more
      coins and profiting from their value.
    sentences:
      - Buyer purchases cryptocurrency from seller in exchange.
      - >-
        Price fluctuates due to fear and uncertainty, only time will reveal its
        direction.
      - >-
        New user's post removed due to lack of required **karma** and account
        age.
  - source_sentence: >-
      User seeks assistance with retrieving funds from a cryptocurrency
      investment platform.
    sentences:
      - Bot removed post for being too short, resubmit with more characters.
      - People enjoy walking while searching for digital currency in their area.
      - >-
        Cryptocurrency project's legitimacy unlikely due to complexity and
        scrutiny in parachain development and ecosystem interactions.
  - source_sentence: >-
      Large cryptocurrencies' market dominance may change as new projects emerge
      with exceptional utility and marketing.
    sentences:
      - Market experiencing significant decline.
      - >-
        Decentralized concept in crypto is main idea, but most coins are
        centralized.
      - Cryptocurrency users share information.
  - source_sentence: Use XLM for low-cost transactions between exchanges, saving on fees.
    sentences:
      - Exchanges should automate process for increased activity.
      - >-
        Investment taxes vary by country, but generally apply after withdrawal,
        with losses still needing declaration.
      - Use basic version, buy coins with credit card.
  - source_sentence: >-
      New user seeks advice on storing Bitcoin and USDT on WazirX or Binance,
      considering pros and cons.
    sentences:
      - >-
        Buy cryptocurrency directly with credit card, but high fee makes Indian
        exchange a better option.
      - 'Cryptocurrency prices: Bitcoin, Ethereum, and others fluctuate.'
      - Investor has faith in Tezos, making strategic moves.

SentenceTransformer based on microsoft/mpnet-base

This is a sentence-transformers model finetuned from microsoft/mpnet-base. 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: microsoft/mpnet-base
Maximum Sequence Length: 512 tokens
Output Dimensionality: 768 tokens
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}) with Transformer model: MPNetModel 
  (1): Pooling({'word_embedding_dimension': 768, '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.

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("pawan2411/crypto_nli")
# Run inference
sentences = [
    'New user seeks advice on storing Bitcoin and USDT on WazirX or Binance, considering pros and cons.',
    'Buy cryptocurrency directly with credit card, but high fee makes Indian exchange a better option.',
    'Investor has faith in Tezos, making strategic moves.',
]
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

Unnamed Dataset

Size: 24,901 training samples
Columns: sentence_0, sentence_1, and label
Approximate statistics based on the first 1000 samples:
sentence_0 sentence_1 label
type string string int
details
min: 6 tokens
mean: 21.86 tokens
max: 61 tokens

min: 6 tokens
mean: 16.67 tokens
max: 50 tokens

0: ~83.50%
1: ~16.50%

	sentence_0	sentence_1	label
type	string	string	int
details	min: 6 tokens mean: 21.86 tokens max: 61 tokens	min: 6 tokens mean: 16.67 tokens max: 50 tokens	0: ~83.50% 1: ~16.50%

Samples:

sentence_0	sentence_1	label
`User asks about tracing crypto swaps and process of exchanging digital currencies.`	`"Private cryptocurrency swap can't be traced."`	`0`
`Cryptocurrency project with weak fundamentals deserves to fail, cherish coins before next market downturn.`	`"Trust information in this community."`	`0`
`New user seeks advice on using crypto credit cards in daily life.`	`User uses digital wallet for cryptocurrency transactions, earning cashback rewards.`	`1`

Loss: SoftmaxLoss

Training Hyperparameters

Non-Default Hyperparameters

per_device_train_batch_size: 64
per_device_eval_batch_size: 64
num_train_epochs: 10
multi_dataset_batch_sampler: round_robin

All Hyperparameters

Click to expand

overwrite_output_dir: False
do_predict: False
eval_strategy: no
prediction_loss_only: True
per_device_train_batch_size: 64
per_device_eval_batch_size: 64
per_gpu_train_batch_size: None
per_gpu_eval_batch_size: None
gradient_accumulation_steps: 1
eval_accumulation_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
num_train_epochs: 10
max_steps: -1
lr_scheduler_type: linear
lr_scheduler_kwargs: {}
warmup_ratio: 0.0
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: 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: 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
batch_sampler: batch_sampler
multi_dataset_batch_sampler: round_robin

Training Logs

Epoch	Step	Training Loss
1.2821	500	0.3912
2.5641	1000	0.3157
3.8462	1500	0.2926
5.1282	2000	0.2788
6.4103	2500	0.2599
7.6923	3000	0.2428
8.9744	3500	0.2314
1.2821	500	0.2333
2.5641	1000	0.2292
3.8462	1500	0.1987
5.1282	2000	0.1757
6.4103	2500	0.1578
7.6923	3000	0.1413
8.9744	3500	0.1258
1.2821	500	0.1086
2.5641	1000	0.1048
3.8462	1500	0.0917
5.1282	2000	0.0805
6.4103	2500	0.0712
7.6923	3000	0.0673
8.9744	3500	0.0646
1.2821	500	0.0505
2.5641	1000	0.0511
3.8462	1500	0.046
5.1282	2000	0.0415
6.4103	2500	0.0396
7.6923	3000	0.0357
8.9744	3500	0.0382
1.2821	500	0.0252
2.5641	1000	0.029
3.8462	1500	0.0247
5.1282	2000	0.0233
6.4103	2500	0.0228
7.6923	3000	0.0218
8.9744	3500	0.0251
1.2821	500	0.0158
2.5641	1000	0.0184
3.8462	1500	0.0165
5.1282	2000	0.0139
6.4103	2500	0.0145
7.6923	3000	0.0139
8.9744	3500	0.0164

Framework Versions

Python: 3.10.12
Sentence Transformers: 3.0.1
Transformers: 4.42.4
PyTorch: 2.3.1+cu121
Accelerate: 0.32.1
Datasets: 2.20.0
Tokenizers: 0.19.1

Citation

BibTeX

Sentence Transformers and SoftmaxLoss

@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",
}