Add new SentenceTransformer model

b6024d8 verified 7 months ago

22.2 kB

language:
  - en
license: apache-2.0
tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - dense
  - generated_from_trainer
  - dataset_size:5600
  - loss:MatryoshkaLoss
  - loss:MultipleNegativesRankingLoss
base_model: Qwen/Qwen3-Embedding-0.6B
widget:
  - source_sentence: What were the total assets at fair value on December 31, 2023?
    sentences:
      - >-
        In addition to its contractual cash requirements, the Company has an
        authorized share repurchase program. The program does not obligate the
        Company to acquire a minimum amount of shares. As of September 30, 2023,
        the Company’s quarterly cash dividend was $0.24 per share.
      - >-
        Effective January 1, 2023, we prospectively adopted new guidance that
        eliminated the recognition and measurement of TDRs. We evaluate all
        loans and receivables restructurings according to accounting guidance
        for loan refinancing and restructuring. Modifications to loans and
        receivables primarily include temporary interest rate reductions and
        placing the customer on a fixed payment plan not to exceed 60 months.
      - >-
        Total assets at fair value on December 31, 2023 were reported to be
        $71,921 million.
  - source_sentence: >-
      What were the key factors affecting the company's cash flow from
      operations in fiscal 2023?
    sentences:
      - >-
        General and administrative | $ | 950 | | $ | 2,025 | 113 | % Percentage
        of revenue | 11 | % | 20 | % | General and administrative expense
        increased $1.1 billion, or 113%, in 2023, compared to 2022.
      - >-
        Within two months after submission of each annual execution proposal,
        the Macao government will decide on their approval, and may request
        adjustments to specific projects, to the investment amount and to the
        execution schedule.
      - >-
        The company's cash flow from operations in fiscal 2023 was affected by
        various factors including changes in working capital components like
        accounts payable, inventories, and accounts receivable.
  - source_sentence: >-
      What percentage of the total U.S. dialysis patient service revenues were
      generated from government-based programs in 2023?
    sentences:
      - >-
        The document includes a 'Glossary of Terms and Acronyms' that provides
        definitions and explanations of financial terms used.
      - >-
        Profit before taxes for 2022 was $8,752 million and rose to $13,050
        million in 2023.
      - >-
        In 2023, approximately 67% of the total U.S. dialysis patient service
        revenues were generated from government-based programs.
  - source_sentence: What was the effective income tax rate for the Company in 2023?
    sentences:
      - >-
        Chevron's oil-equivalent production in the UK has increased by
        approximately 4 percent from 2022 to 2023, as indicated in the
        production summary tables.
      - >-
        The Company’s effective income tax rate decreased to 25.1% in 2023
        compared to 25.9% in the prior year.
      - >-
        In 2023, 45% of our consolidated Gross Merchandise Sales was generated
        when a seller or buyer, or both, were located outside of the United
        States.
  - source_sentence: >-
      Which section of the financial document addresses Financial Statements and
      Supplementary Data?
    sentences:
      - >-
        The gift card liability was $145,014 in 2022 and increased to $164,930
        in 2023.
      - The 7% Notes due 2029 are scheduled to mature on February 15, 2029.
      - >-
        Financial Statements and Supplementary Data are addressed in Item 8 of
        the financial document.
pipeline_tag: sentence-similarity
library_name: sentence-transformers
metrics:
  - cosine_accuracy@1
  - cosine_accuracy@3
  - cosine_accuracy@5
  - cosine_accuracy@10
  - cosine_precision@1
  - cosine_precision@3
  - cosine_precision@5
  - cosine_precision@10
  - cosine_recall@1
  - cosine_recall@3
  - cosine_recall@5
  - cosine_recall@10
  - cosine_ndcg@10
  - cosine_mrr@10
  - cosine_map@100
model-index:
  - name: 'Qwen3 base Financial '
    results:
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 1024
          type: dim_1024
        metrics:
          - type: cosine_accuracy@1
            value: 0.7507142857142857
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.8707142857142857
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.8985714285714286
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.9364285714285714
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.7507142857142857
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.29023809523809524
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.1797142857142857
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.09364285714285712
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.7507142857142857
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.8707142857142857
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.8985714285714286
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.9364285714285714
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.846090041345316
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.8169348072562356
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.8197317550291238
            name: Cosine Map@100

Qwen3 base Financial

This is a sentence-transformers model finetuned from Qwen/Qwen3-Embedding-0.6B on the json dataset. It maps sentences & paragraphs to a 1024-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: Qwen/Qwen3-Embedding-0.6B
Maximum Sequence Length: 32768 tokens
Output Dimensionality: 1024 dimensions
Similarity Function: Cosine Similarity
Training Dataset:
- json
Language: en
License: apache-2.0

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': 32768, 'do_lower_case': False, 'architecture': 'Qwen3Model'})
  (1): Pooling({'word_embedding_dimension': 1024, 'pooling_mode_cls_token': False, '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': True, '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("PhilipCisco/qwen3-base-financial2")
# Run inference
queries = [
    "Which section of the financial document addresses Financial Statements and Supplementary Data?",
]
documents = [
    'Financial Statements and Supplementary Data are addressed in Item 8 of the financial document.',
    'The 7% Notes due 2029 are scheduled to mature on February 15, 2029.',
    'The gift card liability was $145,014 in 2022 and increased to $164,930 in 2023.',
]
query_embeddings = model.encode_query(queries)
document_embeddings = model.encode_document(documents)
print(query_embeddings.shape, document_embeddings.shape)
# [1, 1024] [3, 1024]

# Get the similarity scores for the embeddings
similarities = model.similarity(query_embeddings, document_embeddings)
print(similarities)
# tensor([[0.7375, 0.1121, 0.0035]])

Evaluation

Metrics

Information Retrieval

Dataset: dim_1024
Evaluated with InformationRetrievalEvaluator with these parameters:
```
{
    "truncate_dim": 1024
}
```

Metric	Value
cosine_accuracy@1	0.7507
cosine_accuracy@3	0.8707
cosine_accuracy@5	0.8986
cosine_accuracy@10	0.9364
cosine_precision@1	0.7507
cosine_precision@3	0.2902
cosine_precision@5	0.1797
cosine_precision@10	0.0936
cosine_recall@1	0.7507
cosine_recall@3	0.8707
cosine_recall@5	0.8986
cosine_recall@10	0.9364
cosine_ndcg@10	0.8461
cosine_mrr@10	0.8169
cosine_map@100	0.8197

Training Details

Training Dataset

json

Dataset: json
Size: 5,600 training samples
Columns: anchor and positive
Approximate statistics based on the first 1000 samples:
anchor positive
type string string
details
min: 7 tokens
mean: 20.73 tokens
max: 50 tokens

min: 10 tokens
mean: 47.95 tokens
max: 431 tokens

	anchor	positive
type	string	string
details	min: 7 tokens mean: 20.73 tokens max: 50 tokens	min: 10 tokens mean: 47.95 tokens max: 431 tokens

Samples:

anchor	positive
`What was the increase in sales and marketing expenses for the year ended December 31, 2023 compared to 2022?`	`Sales and marketing expenses increased by $42.5 million, or 6%, for the year ended December 31, 2023 compared to 2022.`
`What method is used to provide information about legal proceedings in the Annual Report on Form 10-K?`	`Information about legal proceedings in the Annual Report on Form 10-K is incorporated by reference under several notes and sections.`
`How did selling, distribution, and administration expenses change in 2023 compared to previous years?`	`In 2023, the decline in Selling, distribution and administration expense was driven by lower compensation expense associated with workforce reductions, lower costs for professional services and lower freight and warehousing expenses as a result of lower shipments during 2023. Additionally, Selling, distribution and administration expense in 2023 included $116.0 million of intangible asset impairment charges as compared to $281.0 million of intangible asset impairment charges in 2022.`

Loss: MatryoshkaLoss with these parameters:

{
    "loss": "MultipleNegativesRankingLoss",
    "matryoshka_dims": [
        1024
    ],
    "matryoshka_weights": [
        1
    ],
    "n_dims_per_step": -1
}

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: epoch
per_device_train_batch_size: 4
per_device_eval_batch_size: 4
gradient_accumulation_steps: 8
learning_rate: 2e-05
num_train_epochs: 4
lr_scheduler_type: cosine
warmup_ratio: 0.1
bf16: True
tf32: 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: epoch
prediction_loss_only: True
per_device_train_batch_size: 4
per_device_eval_batch_size: 4
per_gpu_train_batch_size: None
per_gpu_eval_batch_size: None
gradient_accumulation_steps: 8
eval_accumulation_steps: None
torch_empty_cache_steps: None
learning_rate: 2e-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: cosine
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: True
fp16: False
fp16_opt_level: O1
half_precision_backend: auto
bf16_full_eval: False
fp16_full_eval: False
tf32: True
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}
parallelism_config: 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: 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: no_duplicates
multi_dataset_batch_sampler: proportional
router_mapping: {}
learning_rate_mapping: {}

Training Logs

Epoch	Step	Training Loss	dim_1024_cosine_ndcg@10
-1	-1	-	0.7762
0.1713	10	0.0243	-
0.3426	20	0.0269	-
0.5139	30	0.0171	-
0.6852	40	0.0224	-
0.8565	50	0.0376	-
1.0	59	-	0.8200
1.0171	60	0.0221	-
1.1884	70	0.0089	-
1.3597	80	0.0127	-
1.5310	90	0.0116	-
1.7024	100	0.0086	-
1.8737	110	0.0113	-
2.0	118	-	0.8280
2.0343	120	0.0074	-
2.2056	130	0.0077	-
2.3769	140	0.0107	-
2.5482	150	0.0089	-
2.7195	160	0.0098	-
2.8908	170	0.006	-
3.0	177	-	0.8448
3.0514	180	0.0111	-
3.2227	190	0.0074	-
3.3940	200	0.0082	-
3.5653	210	0.0047	-
3.7366	220	0.0076	-
3.9079	230	0.0085	-
4.0	236	-	0.8461

The bold row denotes the saved checkpoint.

Framework Versions

Python: 3.10.14
Sentence Transformers: 5.1.0
Transformers: 4.56.1
PyTorch: 2.8.0+cu128
Accelerate: 1.10.1
Datasets: 2.19.1
Tokenizers: 0.22.0

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",
}

MatryoshkaLoss

@misc{kusupati2024matryoshka,
    title={Matryoshka Representation Learning},
    author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi},
    year={2024},
    eprint={2205.13147},
    archivePrefix={arXiv},
    primaryClass={cs.LG}
}

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}
}