Update README.md

d2c8b18 verified 10 months ago

27.8 kB

tags:
  - sentence-transformers
  - cross-encoder
  - generated_from_trainer
  - dataset_size:554403
  - loss:BinaryCrossEntropyLoss
base_model: answerdotai/ModernBERT-base
pipeline_tag: text-ranking
library_name: sentence-transformers
metrics:
  - map
  - mrr@10
  - ndcg@10
model-index:
  - name: CrossEncoder based on answerdotai/ModernBERT-base
    results:
      - task:
          type: cross-encoder-reranking
          name: Cross Encoder Reranking
        dataset:
          name: s2orc dev
          type: s2orc-dev
        metrics:
          - type: map
            value: 0.8712
            name: Map
          - type: mrr@10
            value: 0.8711
            name: Mrr@10
          - type: ndcg@10
            value: 0.8765
            name: Ndcg@10
      - task:
          type: cross-encoder-reranking
          name: Cross Encoder Reranking
        dataset:
          name: NanoMSMARCO R100
          type: NanoMSMARCO_R100
        metrics:
          - type: map
            value: 0.4941
            name: Map
          - type: mrr@10
            value: 0.482
            name: Mrr@10
          - type: ndcg@10
            value: 0.5529
            name: Ndcg@10
      - task:
          type: cross-encoder-nano-beir
          name: Cross Encoder Nano BEIR
        dataset:
          name: NanoBEIR R100 mean
          type: NanoBEIR_R100_mean
        metrics:
          - type: map
            value: 0.4941
            name: Map
          - type: mrr@10
            value: 0.482
            name: Mrr@10
          - type: ndcg@10
            value: 0.5529
            name: Ndcg@10

CrossEncoder based on answerdotai/ModernBERT-base

This is a Cross Encoder model finetuned from answerdotai/ModernBERT-base using the sentence-transformers library. It computes scores for pairs of texts, which can be used for text reranking and semantic search.

Model Details

Model Description

Model Type: Cross Encoder
Base model: answerdotai/ModernBERT-base
Maximum Sequence Length: 8192 tokens
Number of Output Labels: 1 label

Model Sources

Documentation: Sentence Transformers Documentation
Documentation: Cross Encoder Documentation
Repository: Sentence Transformers on GitHub
Hugging Face: Cross Encoders on Hugging Face

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 CrossEncoder

# Download from the 🤗 Hub
model = CrossEncoder("Janari01/reranker-ModernBERT-base-s2orc")
# Get scores for pairs of texts
pairs = [
    ["Engineering students' understanding of the role of experimentation", 'Resource constraints have forced engineering schools to reduce laboratory provisions in undergraduate courses. In many instances hands-on experimentation has been replaced by demonstrations or computer simulations. Many engineering educators have cautioned against replacing experiments with simulations on the basis that this will lead to a misunderstanding of the role of experimentation in engineering practice. However, little is known about how students conceptualize the role of experimentation in developing engineering understanding. This study is based on interviews with third-year mechanical engineering students. Findings are presented on their perceptions in relation to the role of experimentation in developing engineering knowledge and practice.'],
    ["Engineering students' understanding of the role of experimentation", '"Excellent engineer training plan"was a core problem for cultivating students\' engineering ability,but at present the students in engineering ability and the enterprise demand disjointed phenomenon had more commons.Based on process equipment and control engineering as an example,for the general undergraduate colleges and universities to cultivate students\' engineering ability and enterprise demand disjointed phenomenon and the existing problems were analyzed,and the relevant approach was put forward,in order to improve students\' engineering ability to provide reference ideas.'],
    ["Engineering students' understanding of the role of experimentation", 'This paper contributes to the discussion of pedagogical training of engineering teachers based on a case study carried out in higher education institutions in Brazil, namely in Electrical Engineering. For this purpose, the authors chose to articulate two research methods: document analysis of the courses offered in the postgraduate programs (Master and PhD) in Electrical Engineering and a survey conducted with students and teachers from 58 of these postgraduate electrical engineering programs. The data analysis indicated that most of the teachers agreed that pedagogical training should be offered to engineering students. Postgraduate students also showed interest in enrolling courses with pedagogic focus. With this analysis we can state that there is a need to rethink engineering education, in order to create conditions for the development of competences related with teaching and learning innovation. This study shows the needs and presents some recommendations to deal with these issues in this field.'],
    ["Engineering students' understanding of the role of experimentation", 'Engineering practical teaching reform in higher institutions centers on improving students’ comprehensive quality,developing their innovative spirit and engineering practice ability,building teaching system for engineering training and demonstration center for engineering training.The article implements practical teaching reform on metalworking practice and electronic practice and provides students with a platform for integrated engineering training,leading them toward competence,quality and innovation development.'],
    ["Engineering students' understanding of the role of experimentation", 'Lisa Benson is an Associate Professor of Engineering and Science Education at Clemson University, with a joint appointment in Bioengineering. Her research focuses on the interactions between student motivation and their learning experiences. Her projects involve the study of student perceptions, beliefs and attitudes towards becoming engineers and scientists, and their problem solving processes. Other projects in the Benson group include effects of student-centered active learning, self-regulated learning, and incorporating engineering into secondary science and mathematics classrooms. Her education includes a B.S. in Bioengineering from the University of Vermont, and M.S. and Ph.D. in Bioengineering from Clemson University.'],
]
scores = model.predict(pairs)
print(scores.shape)
# (5,)

# Or rank different texts based on similarity to a single text
ranks = model.rank(
    "Engineering students' understanding of the role of experimentation",
    [
        'Resource constraints have forced engineering schools to reduce laboratory provisions in undergraduate courses. In many instances hands-on experimentation has been replaced by demonstrations or computer simulations. Many engineering educators have cautioned against replacing experiments with simulations on the basis that this will lead to a misunderstanding of the role of experimentation in engineering practice. However, little is known about how students conceptualize the role of experimentation in developing engineering understanding. This study is based on interviews with third-year mechanical engineering students. Findings are presented on their perceptions in relation to the role of experimentation in developing engineering knowledge and practice.',
        '"Excellent engineer training plan"was a core problem for cultivating students\' engineering ability,but at present the students in engineering ability and the enterprise demand disjointed phenomenon had more commons.Based on process equipment and control engineering as an example,for the general undergraduate colleges and universities to cultivate students\' engineering ability and enterprise demand disjointed phenomenon and the existing problems were analyzed,and the relevant approach was put forward,in order to improve students\' engineering ability to provide reference ideas.',
        'This paper contributes to the discussion of pedagogical training of engineering teachers based on a case study carried out in higher education institutions in Brazil, namely in Electrical Engineering. For this purpose, the authors chose to articulate two research methods: document analysis of the courses offered in the postgraduate programs (Master and PhD) in Electrical Engineering and a survey conducted with students and teachers from 58 of these postgraduate electrical engineering programs. The data analysis indicated that most of the teachers agreed that pedagogical training should be offered to engineering students. Postgraduate students also showed interest in enrolling courses with pedagogic focus. With this analysis we can state that there is a need to rethink engineering education, in order to create conditions for the development of competences related with teaching and learning innovation. This study shows the needs and presents some recommendations to deal with these issues in this field.',
        'Engineering practical teaching reform in higher institutions centers on improving students’ comprehensive quality,developing their innovative spirit and engineering practice ability,building teaching system for engineering training and demonstration center for engineering training.The article implements practical teaching reform on metalworking practice and electronic practice and provides students with a platform for integrated engineering training,leading them toward competence,quality and innovation development.',
        'Lisa Benson is an Associate Professor of Engineering and Science Education at Clemson University, with a joint appointment in Bioengineering. Her research focuses on the interactions between student motivation and their learning experiences. Her projects involve the study of student perceptions, beliefs and attitudes towards becoming engineers and scientists, and their problem solving processes. Other projects in the Benson group include effects of student-centered active learning, self-regulated learning, and incorporating engineering into secondary science and mathematics classrooms. Her education includes a B.S. in Bioengineering from the University of Vermont, and M.S. and Ph.D. in Bioengineering from Clemson University.',
    ]
)
# [{'corpus_id': ..., 'score': ...}, {'corpus_id': ..., 'score': ...}, ...]

Evaluation

Metrics

Cross Encoder Reranking

Dataset: s2orc-dev

Evaluated with CrossEncoderRerankingEvaluator with these parameters:

{
    "at_k": 10,
    "always_rerank_positives": false
}

Metric	Value
map	0.8712 (+0.1333)
mrr@10	0.8711 (+0.1351)
ndcg@10	0.8765 (+0.1106)

Cross Encoder Reranking

Dataset: NanoMSMARCO_R100

Evaluated with CrossEncoderRerankingEvaluator with these parameters:

{
    "at_k": 10,
    "always_rerank_positives": true
}

Metric	Value
map	0.4941 (+0.0045)
mrr@10	0.4820 (+0.0045)
ndcg@10	0.5529 (+0.0124)

Cross Encoder Nano BEIR

Dataset: NanoBEIR_R100_mean

Evaluated with CrossEncoderNanoBEIREvaluator with these parameters:

{
    "dataset_names": [
        "msmarco"
    ],
    "rerank_k": 100,
    "at_k": 10,
    "always_rerank_positives": true
}

Metric	Value
map	0.4941 (+0.0045)
mrr@10	0.4820 (+0.0045)
ndcg@10	0.5529 (+0.0124)

Training Details

Training Dataset

Unnamed Dataset

Size: 554,403 training samples
Columns: title, abstract, and label

Approximate statistics based on the first 1000 samples:

	title	abstract	label
type	string	string	int
details	min: 33 characters mean: 83.77 characters max: 162 characters	min: 91 characters mean: 678.94 characters max: 1023 characters	0: ~81.80% 1: ~18.20%

Samples:

title	abstract	label
`Engineering students' understanding of the role of experimentation`	Resource constraints have forced engineering schools to reduce laboratory provisions in undergraduate courses. In many instances hands-on experimentation has been replaced by demonstrations or computer simulations. Many engineering educators have cautioned against replacing experiments with simulations on the basis that this will lead to a misunderstanding of the role of experimentation in engineering practice. However, little is known about how students conceptualize the role of experimentation in developing engineering understanding. This study is based on interviews with third-year mechanical engineering students. Findings are presented on their perceptions in relation to the role of experimentation in developing engineering knowledge and practice.	`1`
`Engineering students' understanding of the role of experimentation`	"Excellent engineer training plan"was a core problem for cultivating students' engineering ability,but at present the students in engineering ability and the enterprise demand disjointed phenomenon had more commons.Based on process equipment and control engineering as an example,for the general undergraduate colleges and universities to cultivate students' engineering ability and enterprise demand disjointed phenomenon and the existing problems were analyzed,and the relevant approach was put forward,in order to improve students' engineering ability to provide reference ideas.	`0`
`Engineering students' understanding of the role of experimentation`	This paper contributes to the discussion of pedagogical training of engineering teachers based on a case study carried out in higher education institutions in Brazil, namely in Electrical Engineering. For this purpose, the authors chose to articulate two research methods: document analysis of the courses offered in the postgraduate programs (Master and PhD) in Electrical Engineering and a survey conducted with students and teachers from 58 of these postgraduate electrical engineering programs. The data analysis indicated that most of the teachers agreed that pedagogical training should be offered to engineering students. Postgraduate students also showed interest in enrolling courses with pedagogic focus. With this analysis we can state that there is a need to rethink engineering education, in order to create conditions for the development of competences related with teaching and learning innovation. This study shows the needs and presents some recommendations to deal with these issues...	`0`

Loss: BinaryCrossEntropyLoss with these parameters:

{
    "activation_fn": "torch.nn.modules.linear.Identity",
    "pos_weight": 5
}

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: steps
per_device_train_batch_size: 16
per_device_eval_batch_size: 16
learning_rate: 2e-05
num_train_epochs: 1
warmup_ratio: 0.1
seed: 12
bf16: True
dataloader_num_workers: 6
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: 16
per_device_eval_batch_size: 16
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: 0.0
adam_beta1: 0.9
adam_beta2: 0.999
adam_epsilon: 1e-08
max_grad_norm: 1.0
num_train_epochs: 1
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: 12
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: None
local_rank: 0
ddp_backend: None
tpu_num_cores: None
tpu_metrics_debug: False
debug: []
dataloader_drop_last: False
dataloader_num_workers: 6
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
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
eval_use_gather_object: False
average_tokens_across_devices: False
prompts: None
batch_sampler: batch_sampler
multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	s2orc-dev_ndcg@10	NanoMSMARCO_R100_ndcg@10	NanoBEIR_R100_mean_ndcg@10
-1	-1	-	0.1165 (-0.6495)	0.0426 (-0.4978)	0.0426 (-0.4978)
0.0000	1	1.0682	-	-	-
0.0144	500	1.1555	-	-	-
0.0289	1000	0.7743	-	-	-
0.0433	1500	0.538	-	-	-
0.0577	2000	0.5771	-	-	-
0.0721	2500	0.5345	-	-	-
0.0866	3000	0.4394	-	-	-
0.1010	3500	0.4607	-	-	-
0.1154	4000	0.3866	0.8685 (+0.1025)	0.5469 (+0.0064)	0.5469 (+0.0064)
0.1299	4500	0.4222	-	-	-
0.1443	5000	0.3734	-	-	-
0.1587	5500	0.3558	-	-	-
0.1732	6000	0.3968	-	-	-
0.1876	6500	0.3203	-	-	-
0.2020	7000	0.3354	-	-	-
0.2164	7500	0.3579	-	-	-
0.2309	8000	0.3349	0.8765 (+0.1106)	0.5529 (+0.0124)	0.5529 (+0.0124)

Framework Versions

Python: 3.9.13
Sentence Transformers: 4.1.0
Transformers: 4.52.4
PyTorch: 2.7.1+cu118
Accelerate: 1.7.0
Datasets: 3.6.0
Tokenizers: 0.21.1

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",
}