--- language: - en license: apache-2.0 tags: - sentence-transformers - cross-encoder - generated_from_trainer - dataset_size:2749365 - loss:BinaryCrossEntropyLoss base_model: answerdotai/ModernBERT-base pipeline_tag: text-ranking library_name: sentence-transformers metrics: - map - mrr@10 - ndcg@10 model-index: - name: ModernBERT-base trained on GooAQ results: - task: type: cross-encoder-reranking name: Cross Encoder Reranking dataset: name: gooaq dev type: gooaq-dev metrics: - type: map value: 0.5439 name: Map - type: mrr@10 value: 0.5411 name: Mrr@10 - type: ndcg@10 value: 0.5936 name: Ndcg@10 - task: type: cross-encoder-reranking name: Cross Encoder Reranking dataset: name: NanoMSMARCO R100 type: NanoMSMARCO_R100 metrics: - type: map value: 0.3929 name: Map - type: mrr@10 value: 0.3751 name: Mrr@10 - type: ndcg@10 value: 0.4428 name: Ndcg@10 - task: type: cross-encoder-reranking name: Cross Encoder Reranking dataset: name: NanoNFCorpus R100 type: NanoNFCorpus_R100 metrics: - type: map value: 0.3119 name: Map - type: mrr@10 value: 0.4287 name: Mrr@10 - type: ndcg@10 value: 0.314 name: Ndcg@10 - task: type: cross-encoder-reranking name: Cross Encoder Reranking dataset: name: NanoNQ R100 type: NanoNQ_R100 metrics: - type: map value: 0.3869 name: Map - type: mrr@10 value: 0.3837 name: Mrr@10 - type: ndcg@10 value: 0.4316 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: 0.3639 name: Map - type: mrr@10 value: 0.3959 name: Mrr@10 - type: ndcg@10 value: 0.3961 name: Ndcg@10 --- # ModernBERT-base trained on GooAQ This is a [Cross Encoder](https://www.sbert.net/docs/cross_encoder/usage/usage.html) model finetuned from [answerdotai/ModernBERT-base](https://huggingface.co/answerdotai/ModernBERT-base) using the [sentence-transformers](https://www.SBERT.net) 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](https://huggingface.co/answerdotai/ModernBERT-base) - **Maximum Sequence Length:** 8192 tokens - **Number of Output Labels:** 1 label - **Language:** en - **License:** apache-2.0 ### Model Sources - **Documentation:** [Sentence Transformers Documentation](https://sbert.net) - **Documentation:** [Cross Encoder Documentation](https://www.sbert.net/docs/cross_encoder/usage/usage.html) - **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers) - **Hugging Face:** [Cross Encoders on Hugging Face](https://huggingface.co/models?library=sentence-transformers&other=cross-encoder) ## Usage ### Direct Usage (Sentence Transformers) First install the Sentence Transformers library: ```bash pip install -U sentence-transformers ``` Then you can load this model and run inference. ```python from sentence_transformers import CrossEncoder # Download from the 🤗 Hub model = CrossEncoder("ayushexel/ce-modernbert-trained-1epoch") # Get scores for pairs of texts pairs = [ ['can you still get pregnant if you are infertile?', 'Many infertile couples will go on to conceive a child without treatment. After trying to get pregnant for two years, about 95 percent of couples successfully conceive.'], ['can you still get pregnant if you are infertile?', 'Secondary infertility is the inability to become pregnant or to carry a baby to term after previously giving birth to a baby. Secondary infertility shares many of the same causes of primary infertility. Secondary infertility might be caused by: Impaired sperm production, function or delivery in men.'], ['can you still get pregnant if you are infertile?', "Problems with cervical mucus can interfere with getting pregnant. Mild cases may increase the time it takes to get pregnant, but won't necessarily cause infertility."], ['can you still get pregnant if you are infertile?', 'No treatment can stop the process of diminished ovarian reserve, but women who are infertile due to low egg count or quality can sometimes use assisted reproductive technologies to achieve a pregnancy.'], ['can you still get pregnant if you are infertile?', "Human conception requires an egg and sperm. If you're not ovulating, you won't be able to get pregnant. Anovulation is a common cause of female infertility and it can be triggered by many conditions. Most women who are experiencing ovulation problems have irregular periods."], ] scores = model.predict(pairs) print(scores.shape) # (5,) # Or rank different texts based on similarity to a single text ranks = model.rank( 'can you still get pregnant if you are infertile?', [ 'Many infertile couples will go on to conceive a child without treatment. After trying to get pregnant for two years, about 95 percent of couples successfully conceive.', 'Secondary infertility is the inability to become pregnant or to carry a baby to term after previously giving birth to a baby. Secondary infertility shares many of the same causes of primary infertility. Secondary infertility might be caused by: Impaired sperm production, function or delivery in men.', "Problems with cervical mucus can interfere with getting pregnant. Mild cases may increase the time it takes to get pregnant, but won't necessarily cause infertility.", 'No treatment can stop the process of diminished ovarian reserve, but women who are infertile due to low egg count or quality can sometimes use assisted reproductive technologies to achieve a pregnancy.', "Human conception requires an egg and sperm. If you're not ovulating, you won't be able to get pregnant. Anovulation is a common cause of female infertility and it can be triggered by many conditions. Most women who are experiencing ovulation problems have irregular periods.", ] ) # [{'corpus_id': ..., 'score': ...}, {'corpus_id': ..., 'score': ...}, ...] ``` ## Evaluation ### Metrics #### Cross Encoder Reranking * Dataset: `gooaq-dev` * Evaluated with [CrossEncoderRerankingEvaluator](https://sbert.net/docs/package_reference/cross_encoder/evaluation.html#sentence_transformers.cross_encoder.evaluation.CrossEncoderRerankingEvaluator) with these parameters: ```json { "at_k": 10, "always_rerank_positives": false } ``` | Metric | Value | |:------------|:---------------------| | map | 0.5439 (+0.1636) | | mrr@10 | 0.5411 (+0.1708) | | **ndcg@10** | **0.5936 (+0.1609)** | #### Cross Encoder Reranking * Datasets: `NanoMSMARCO_R100`, `NanoNFCorpus_R100` and `NanoNQ_R100` * Evaluated with [CrossEncoderRerankingEvaluator](https://sbert.net/docs/package_reference/cross_encoder/evaluation.html#sentence_transformers.cross_encoder.evaluation.CrossEncoderRerankingEvaluator) with these parameters: ```json { "at_k": 10, "always_rerank_positives": true } ``` | Metric | NanoMSMARCO_R100 | NanoNFCorpus_R100 | NanoNQ_R100 | |:------------|:---------------------|:---------------------|:---------------------| | map | 0.3929 (-0.0967) | 0.3119 (+0.0509) | 0.3869 (-0.0327) | | mrr@10 | 0.3751 (-0.1024) | 0.4287 (-0.0711) | 0.3837 (-0.0429) | | **ndcg@10** | **0.4428 (-0.0976)** | **0.3140 (-0.0110)** | **0.4316 (-0.0691)** | #### Cross Encoder Nano BEIR * Dataset: `NanoBEIR_R100_mean` * Evaluated with [CrossEncoderNanoBEIREvaluator](https://sbert.net/docs/package_reference/cross_encoder/evaluation.html#sentence_transformers.cross_encoder.evaluation.CrossEncoderNanoBEIREvaluator) with these parameters: ```json { "dataset_names": [ "msmarco", "nfcorpus", "nq" ], "rerank_k": 100, "at_k": 10, "always_rerank_positives": true } ``` | Metric | Value | |:------------|:---------------------| | map | 0.3639 (-0.0261) | | mrr@10 | 0.3959 (-0.0722) | | **ndcg@10** | **0.3961 (-0.0592)** | ## Training Details ### Training Dataset #### Unnamed Dataset * Size: 2,749,365 training samples * Columns: question, answer, and label * Approximate statistics based on the first 1000 samples: | | question | answer | label | |:--------|:-----------------------------------------------------------------------------------------------|:-------------------------------------------------------------------------------------------------|:------------------------------------------------| | type | string | string | int | | details | | | | * Samples: | question | answer | label | |:--------------------------------------------------------------|:--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------| | can you still get pregnant if you are infertile? | Many infertile couples will go on to conceive a child without treatment. After trying to get pregnant for two years, about 95 percent of couples successfully conceive. | 1 | | can you still get pregnant if you are infertile? | Secondary infertility is the inability to become pregnant or to carry a baby to term after previously giving birth to a baby. Secondary infertility shares many of the same causes of primary infertility. Secondary infertility might be caused by: Impaired sperm production, function or delivery in men. | 0 | | can you still get pregnant if you are infertile? | Problems with cervical mucus can interfere with getting pregnant. Mild cases may increase the time it takes to get pregnant, but won't necessarily cause infertility. | 0 | * Loss: [BinaryCrossEntropyLoss](https://sbert.net/docs/package_reference/cross_encoder/losses.html#binarycrossentropyloss) with these parameters: ```json { "activation_fn": "torch.nn.modules.linear.Identity", "pos_weight": 5 } ``` ### Training Hyperparameters #### Non-Default Hyperparameters - `eval_strategy`: steps - `per_device_train_batch_size`: 256 - `per_device_eval_batch_size`: 256 - `learning_rate`: 2e-05 - `num_train_epochs`: 1 - `warmup_ratio`: 0.1 - `seed`: 12 - `bf16`: True - `dataloader_num_workers`: 12 - `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`: 256 - `per_device_eval_batch_size`: 256 - `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`: 12 - `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} - `tp_size`: 0 - `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 - `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 - `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 | gooaq-dev_ndcg@10 | NanoMSMARCO_R100_ndcg@10 | NanoNFCorpus_R100_ndcg@10 | NanoNQ_R100_ndcg@10 | NanoBEIR_R100_mean_ndcg@10 | |:------:|:----:|:-------------:|:-----------------:|:------------------------:|:-------------------------:|:-------------------:|:--------------------------:| | -1 | -1 | - | 0.1022 (-0.3304) | 0.0716 (-0.4688) | 0.2417 (-0.0833) | 0.0286 (-0.4720) | 0.1140 (-0.3414) | | 0.0001 | 1 | 1.3449 | - | - | - | - | - | | 0.0186 | 200 | 1.2174 | - | - | - | - | - | | 0.0372 | 400 | 1.156 | - | - | - | - | - | | 0.0559 | 600 | 0.8504 | - | - | - | - | - | | 0.0745 | 800 | 0.7192 | - | - | - | - | - | | 0.0931 | 1000 | 0.6675 | 0.5936 (+0.1609) | 0.4428 (-0.0976) | 0.3140 (-0.0110) | 0.4316 (-0.0691) | 0.3961 (-0.0592) | ### Framework Versions - Python: 3.11.0 - Sentence Transformers: 4.0.1 - Transformers: 4.50.3 - PyTorch: 2.6.0+cu124 - Accelerate: 1.5.2 - Datasets: 3.5.0 - Tokenizers: 0.21.1 ## Citation ### BibTeX #### Sentence Transformers ```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", } ```