fine_tuned_model_8 / README.md

srikarvar

Add new SentenceTransformer model.

7e4dd74 verified over 1 year ago

preview code

raw

history blame contribute delete

27.4 kB

metadata

base_model: intfloat/multilingual-e5-small
library_name: sentence-transformers
metrics:
  - cosine_accuracy
  - cosine_accuracy_threshold
  - cosine_f1
  - cosine_f1_threshold
  - cosine_precision
  - cosine_recall
  - cosine_ap
  - dot_accuracy
  - dot_accuracy_threshold
  - dot_f1
  - dot_f1_threshold
  - dot_precision
  - dot_recall
  - dot_ap
  - manhattan_accuracy
  - manhattan_accuracy_threshold
  - manhattan_f1
  - manhattan_f1_threshold
  - manhattan_precision
  - manhattan_recall
  - manhattan_ap
  - euclidean_accuracy
  - euclidean_accuracy_threshold
  - euclidean_f1
  - euclidean_f1_threshold
  - euclidean_precision
  - euclidean_recall
  - euclidean_ap
  - max_accuracy
  - max_accuracy_threshold
  - max_f1
  - max_f1_threshold
  - max_precision
  - max_recall
  - max_ap
pipeline_tag: sentence-similarity
tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:2305
  - loss:OnlineContrastiveLoss
widget:
  - source_sentence: Steps to start a vegetable garden
    sentences:
      - How to plant a vegetable garden?
      - If there will be a war between India and Pakistan who will win?
      - What is the most visited tourist attraction in the world?
  - source_sentence: What's the best way to jump rope?
    sentences:
      - If I jump rope for five minutes, how many calories will I use?
      - >-
        You can collaborate on models and datasets using Machine Learning
        platforms by joining the community and accessing enhanced resources.
      - How can I improve my public speaking skills?
  - source_sentence: How can remote team management be improved?
    sentences:
      - What are the key challenges of managing remote teams?
      - The library supports various audio formats such as WAV, MP3, and FLAC.
      - >-
        The `validate_data` method is used to perform checks on the data set for
        correctness.
  - source_sentence: Latest advancements in quantum computing
    sentences:
      - How to cook a turkey?
      - Latest advancements in AI
      - How to create a resume?
  - source_sentence: >-
      Practical guides are available to assist you in achieving specific goals
      and addressing real-world challenges with the framework.
    sentences:
      - How to bake cookies?
      - >-
        Yes, there are practical guides to help you achieve specific objectives
        and solve real-world problems with the framework.
      - How to create an email signature?
model-index:
  - name: SentenceTransformer based on intfloat/multilingual-e5-small
    results:
      - task:
          type: binary-classification
          name: Binary Classification
        dataset:
          name: pair class dev
          type: pair-class-dev
        metrics:
          - type: cosine_accuracy
            value: 0.9182879377431906
            name: Cosine Accuracy
          - type: cosine_accuracy_threshold
            value: 0.8421422243118286
            name: Cosine Accuracy Threshold
          - type: cosine_f1
            value: 0.920754716981132
            name: Cosine F1
          - type: cosine_f1_threshold
            value: 0.8421422243118286
            name: Cosine F1 Threshold
          - type: cosine_precision
            value: 0.9037037037037037
            name: Cosine Precision
          - type: cosine_recall
            value: 0.9384615384615385
            name: Cosine Recall
          - type: cosine_ap
            value: 0.9452952670187734
            name: Cosine Ap
          - type: dot_accuracy
            value: 0.9182879377431906
            name: Dot Accuracy
          - type: dot_accuracy_threshold
            value: 0.8421421647071838
            name: Dot Accuracy Threshold
          - type: dot_f1
            value: 0.920754716981132
            name: Dot F1
          - type: dot_f1_threshold
            value: 0.8421421647071838
            name: Dot F1 Threshold
          - type: dot_precision
            value: 0.9037037037037037
            name: Dot Precision
          - type: dot_recall
            value: 0.9384615384615385
            name: Dot Recall
          - type: dot_ap
            value: 0.9452952670187734
            name: Dot Ap
          - type: manhattan_accuracy
            value: 0.9182879377431906
            name: Manhattan Accuracy
          - type: manhattan_accuracy_threshold
            value: 8.50709342956543
            name: Manhattan Accuracy Threshold
          - type: manhattan_f1
            value: 0.9195402298850576
            name: Manhattan F1
          - type: manhattan_f1_threshold
            value: 8.64261245727539
            name: Manhattan F1 Threshold
          - type: manhattan_precision
            value: 0.916030534351145
            name: Manhattan Precision
          - type: manhattan_recall
            value: 0.9230769230769231
            name: Manhattan Recall
          - type: manhattan_ap
            value: 0.9453829621939649
            name: Manhattan Ap
          - type: euclidean_accuracy
            value: 0.9182879377431906
            name: Euclidean Accuracy
          - type: euclidean_accuracy_threshold
            value: 0.5618841648101807
            name: Euclidean Accuracy Threshold
          - type: euclidean_f1
            value: 0.920754716981132
            name: Euclidean F1
          - type: euclidean_f1_threshold
            value: 0.5618841648101807
            name: Euclidean F1 Threshold
          - type: euclidean_precision
            value: 0.9037037037037037
            name: Euclidean Precision
          - type: euclidean_recall
            value: 0.9384615384615385
            name: Euclidean Recall
          - type: euclidean_ap
            value: 0.9452952670187734
            name: Euclidean Ap
          - type: max_accuracy
            value: 0.9182879377431906
            name: Max Accuracy
          - type: max_accuracy_threshold
            value: 8.50709342956543
            name: Max Accuracy Threshold
          - type: max_f1
            value: 0.920754716981132
            name: Max F1
          - type: max_f1_threshold
            value: 8.64261245727539
            name: Max F1 Threshold
          - type: max_precision
            value: 0.916030534351145
            name: Max Precision
          - type: max_recall
            value: 0.9384615384615385
            name: Max Recall
          - type: max_ap
            value: 0.9453829621939649
            name: Max Ap
      - task:
          type: binary-classification
          name: Binary Classification
        dataset:
          name: pair class test
          type: pair-class-test
        metrics:
          - type: cosine_accuracy
            value: 0.9182879377431906
            name: Cosine Accuracy
          - type: cosine_accuracy_threshold
            value: 0.8421422243118286
            name: Cosine Accuracy Threshold
          - type: cosine_f1
            value: 0.920754716981132
            name: Cosine F1
          - type: cosine_f1_threshold
            value: 0.8421422243118286
            name: Cosine F1 Threshold
          - type: cosine_precision
            value: 0.9037037037037037
            name: Cosine Precision
          - type: cosine_recall
            value: 0.9384615384615385
            name: Cosine Recall
          - type: cosine_ap
            value: 0.9452952670187734
            name: Cosine Ap
          - type: dot_accuracy
            value: 0.9182879377431906
            name: Dot Accuracy
          - type: dot_accuracy_threshold
            value: 0.8421421647071838
            name: Dot Accuracy Threshold
          - type: dot_f1
            value: 0.920754716981132
            name: Dot F1
          - type: dot_f1_threshold
            value: 0.8421421647071838
            name: Dot F1 Threshold
          - type: dot_precision
            value: 0.9037037037037037
            name: Dot Precision
          - type: dot_recall
            value: 0.9384615384615385
            name: Dot Recall
          - type: dot_ap
            value: 0.9452952670187734
            name: Dot Ap
          - type: manhattan_accuracy
            value: 0.9182879377431906
            name: Manhattan Accuracy
          - type: manhattan_accuracy_threshold
            value: 8.50709342956543
            name: Manhattan Accuracy Threshold
          - type: manhattan_f1
            value: 0.9195402298850576
            name: Manhattan F1
          - type: manhattan_f1_threshold
            value: 8.64261245727539
            name: Manhattan F1 Threshold
          - type: manhattan_precision
            value: 0.916030534351145
            name: Manhattan Precision
          - type: manhattan_recall
            value: 0.9230769230769231
            name: Manhattan Recall
          - type: manhattan_ap
            value: 0.9453829621939649
            name: Manhattan Ap
          - type: euclidean_accuracy
            value: 0.9182879377431906
            name: Euclidean Accuracy
          - type: euclidean_accuracy_threshold
            value: 0.5618841648101807
            name: Euclidean Accuracy Threshold
          - type: euclidean_f1
            value: 0.920754716981132
            name: Euclidean F1
          - type: euclidean_f1_threshold
            value: 0.5618841648101807
            name: Euclidean F1 Threshold
          - type: euclidean_precision
            value: 0.9037037037037037
            name: Euclidean Precision
          - type: euclidean_recall
            value: 0.9384615384615385
            name: Euclidean Recall
          - type: euclidean_ap
            value: 0.9452952670187734
            name: Euclidean Ap
          - type: max_accuracy
            value: 0.9182879377431906
            name: Max Accuracy
          - type: max_accuracy_threshold
            value: 8.50709342956543
            name: Max Accuracy Threshold
          - type: max_f1
            value: 0.920754716981132
            name: Max F1
          - type: max_f1_threshold
            value: 8.64261245727539
            name: Max F1 Threshold
          - type: max_precision
            value: 0.916030534351145
            name: Max Precision
          - type: max_recall
            value: 0.9384615384615385
            name: Max Recall
          - type: max_ap
            value: 0.9453829621939649
            name: Max Ap

SentenceTransformer based on intfloat/multilingual-e5-small

This is a sentence-transformers model finetuned from intfloat/multilingual-e5-small. 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: intfloat/multilingual-e5-small
Maximum Sequence Length: 512 tokens
Output Dimensionality: 384 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: BertModel 
  (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})
  (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("srikarvar/fine_tuned_model_8")
# Run inference
sentences = [
    'Practical guides are available to assist you in achieving specific goals and addressing real-world challenges with the framework.',
    'Yes, there are practical guides to help you achieve specific objectives and solve real-world problems with the framework.',
    'How to bake cookies?',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 384]

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

Evaluation

Metrics

Binary Classification

Dataset: pair-class-dev
Evaluated with BinaryClassificationEvaluator

Metric	Value
cosine_accuracy	0.9183
cosine_accuracy_threshold	0.8421
cosine_f1	0.9208
cosine_f1_threshold	0.8421
cosine_precision	0.9037
cosine_recall	0.9385
cosine_ap	0.9453
dot_accuracy	0.9183
dot_accuracy_threshold	0.8421
dot_f1	0.9208
dot_f1_threshold	0.8421
dot_precision	0.9037
dot_recall	0.9385
dot_ap	0.9453
manhattan_accuracy	0.9183
manhattan_accuracy_threshold	8.5071
manhattan_f1	0.9195
manhattan_f1_threshold	8.6426
manhattan_precision	0.916
manhattan_recall	0.9231
manhattan_ap	0.9454
euclidean_accuracy	0.9183
euclidean_accuracy_threshold	0.5619
euclidean_f1	0.9208
euclidean_f1_threshold	0.5619
euclidean_precision	0.9037
euclidean_recall	0.9385
euclidean_ap	0.9453
max_accuracy	0.9183
max_accuracy_threshold	8.5071
max_f1	0.9208
max_f1_threshold	8.6426
max_precision	0.916
max_recall	0.9385
max_ap	0.9454

Binary Classification

Dataset: pair-class-test
Evaluated with BinaryClassificationEvaluator

Metric	Value
cosine_accuracy	0.9183
cosine_accuracy_threshold	0.8421
cosine_f1	0.9208
cosine_f1_threshold	0.8421
cosine_precision	0.9037
cosine_recall	0.9385
cosine_ap	0.9453
dot_accuracy	0.9183
dot_accuracy_threshold	0.8421
dot_f1	0.9208
dot_f1_threshold	0.8421
dot_precision	0.9037
dot_recall	0.9385
dot_ap	0.9453
manhattan_accuracy	0.9183
manhattan_accuracy_threshold	8.5071
manhattan_f1	0.9195
manhattan_f1_threshold	8.6426
manhattan_precision	0.916
manhattan_recall	0.9231
manhattan_ap	0.9454
euclidean_accuracy	0.9183
euclidean_accuracy_threshold	0.5619
euclidean_f1	0.9208
euclidean_f1_threshold	0.5619
euclidean_precision	0.9037
euclidean_recall	0.9385
euclidean_ap	0.9453
max_accuracy	0.9183
max_accuracy_threshold	8.5071
max_f1	0.9208
max_f1_threshold	8.6426
max_precision	0.916
max_recall	0.9385
max_ap	0.9454

Training Details

Training Dataset

Unnamed Dataset

Size: 2,305 training samples
Columns: sentence2, sentence1, and label
Approximate statistics based on the first 1000 samples:
sentence2 sentence1 label
type string string int
details
min: 5 tokens
mean: 13.74 tokens
max: 54 tokens

min: 6 tokens
mean: 14.13 tokens
max: 66 tokens

0: ~43.00%
1: ~57.00%

	sentence2	sentence1	label
type	string	string	int
details	min: 5 tokens mean: 13.74 tokens max: 54 tokens	min: 6 tokens mean: 14.13 tokens max: 66 tokens	0: ~43.00% 1: ~57.00%

Samples:

sentence2	sentence1	label
`What are the components of a computer?`	`How does a computer work?`	`0`
`You have the option to create your own personal blog with the help of Blogging Platforms.`	`Yes, you can start your own personal blog using Blogging Platforms.`	`1`
`It provides the layout of the data and its components.`	`It returns the structure of the data and its fields.`	`1`

Loss: OnlineContrastiveLoss

Evaluation Dataset

Unnamed Dataset

Size: 257 evaluation samples
Columns: sentence2, sentence1, and label
Approximate statistics based on the first 257 samples:
sentence2 sentence1 label
type string string int
details
min: 4 tokens
mean: 14.92 tokens
max: 50 tokens

min: 6 tokens
mean: 14.84 tokens
max: 51 tokens

0: ~49.42%
1: ~50.58%

	sentence2	sentence1	label
type	string	string	int
details	min: 4 tokens mean: 14.92 tokens max: 50 tokens	min: 6 tokens mean: 14.84 tokens max: 51 tokens	0: ~49.42% 1: ~50.58%

Samples:

sentence2	sentence1	label
`What is the speed of sound in air?`	`What is the speed of light in a vacuum?`	`0`
`Steps to fix a leaking faucet`	`How to repair a leaking faucet?`	`1`
`Total bones in an adult human`	`How many bones are in the human body?`	`1`

Loss: OnlineContrastiveLoss

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: epoch
per_device_train_batch_size: 32
per_device_eval_batch_size: 32
gradient_accumulation_steps: 2
num_train_epochs: 4
warmup_ratio: 0.1
load_best_model_at_end: True
optim: adamw_torch_fused
batch_sampler: no_duplicates

All Hyperparameters

Click to expand

overwrite_output_dir: False
do_predict: False
eval_strategy: epoch
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: 2
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: 1.0
num_train_epochs: 4
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: 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_fused
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
batch_sampler: no_duplicates
multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	loss	pair-class-dev_max_ap	pair-class-test_max_ap
0	0	-	-	0.7947	-
0.2740	10	1.6052	-	-	-
0.5479	20	0.8914	-	-	-
0.8219	30	0.8434	-	-	-
0.9863	36	-	0.6144	0.9366	-
1.0959	40	0.7351	-	-	-
1.3699	50	0.5016	-	-	-
1.6438	60	0.3754	-	-	-
1.9178	70	0.3364	-	-	-
2.0	73	-	0.5985	0.9396	-
2.1918	80	0.3456	-	-	-
2.4658	90	0.1953	-	-	-
2.7397	100	0.1186	-	-	-
2.9863	109	-	0.5853	0.9455	-
3.0137	110	0.1622	-	-	-
3.2877	120	0.1863	-	-	-
3.5616	130	0.0906	-	-	-
3.8356	140	0.1035	-	-	-
3.9452	144	-	0.5461	0.9454	0.9454

The bold row denotes the saved checkpoint.

Framework Versions

Python: 3.10.12
Sentence Transformers: 3.1.0
Transformers: 4.41.2
PyTorch: 2.1.2+cu121
Accelerate: 0.34.2
Datasets: 2.19.1
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",
}