Token Classification
Transformers
Safetensors
English
distilbert
named-entity-recognition
ner
bio
Eval Results (legacy)
Instructions to use Hasanblanka/polygraf-ner-stage3-first-model with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Hasanblanka/polygraf-ner-stage3-first-model with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("token-classification", model="Hasanblanka/polygraf-ner-stage3-first-model")# Load model directly from transformers import AutoTokenizer, AutoModelForTokenClassification tokenizer = AutoTokenizer.from_pretrained("Hasanblanka/polygraf-ner-stage3-first-model") model = AutoModelForTokenClassification.from_pretrained("Hasanblanka/polygraf-ner-stage3-first-model", device_map="auto") - Notebooks
- Google Colab
- Kaggle
| stage: 3 | |
| status: completed | |
| dataset: | |
| repo_id: Hasanblanka/polygraf-ner-stage_2-dataset | |
| revision: bce1f32778a333297ba1bf9749c13dd394e6256d | |
| train_split: train | |
| test_split: test | |
| expected_train_records: 506 | |
| expected_test_records: 127 | |
| model: | |
| repo_id: distilbert/distilbert-base-cased | |
| revision: 6ea81172465e8b0ad3fddeed32b986cdcdcffcf0 | |
| architecture: DistilBertForTokenClassification | |
| parameter_count_base: 65800000 | |
| hidden_size: 768 | |
| hidden_layers: 6 | |
| attention_heads: 12 | |
| license: apache-2.0 | |
| reproducibility: | |
| seed: 42 | |
| validation_split: false | |
| local_environment: machinelearning | |
| device: cpu | |
| deterministic_algorithms: true | |
| data_preparation: | |
| max_length: 512 | |
| truncation: false | |
| dynamic_padding: true | |
| special_token_label_id: -100 | |
| prepared_dir: data/interim/stage3_bio | |
| summary_path: artifacts/stage3/bio_preparation_summary.json | |
| training: | |
| epochs: 4 | |
| train_batch_size: 8 | |
| eval_batch_size: 8 | |
| gradient_accumulation_steps: 2 | |
| effective_batch_size: 16 | |
| learning_rate: 0.00005 | |
| weight_decay: 0.01 | |
| warmup_ratio: 0.1 | |
| max_grad_norm: 1.0 | |
| logging_steps: 10 | |
| cpu_threads: 8 | |
| num_workers: 0 | |
| benchmark: | |
| batch_size: 1 | |
| warmup_batches: 5 | |
| measured_batches: 100 | |
| outputs: | |
| model_dir: models/stage3_first_model | |
| artifact_dir: artifacts/stage3 | |
| train_predictions: artifacts/stage3/train_predictions.jsonl | |
| test_predictions: artifacts/stage3/test_predictions.jsonl | |
| training_summary: artifacts/stage3/training_summary.json | |
| training_history: artifacts/stage3/training_history.json | |
| train_metrics: artifacts/stage3/train_metrics.json | |
| test_metrics: artifacts/stage3/test_metrics.json | |
| inference_benchmark: artifacts/stage3/inference_benchmark.json | |
| labels: | |
| entity_order: | |
| - PERSON | |
| - ORGANIZATION | |
| - LOCATION | |
| - TIMEDATE | |
| - PRODUCT | |
| - WORKOFART | |
| - JOB | |
| - AMOUNT | |
| bio_labels: | |
| - O | |
| - B-PERSON | |
| - I-PERSON | |
| - B-ORGANIZATION | |
| - I-ORGANIZATION | |
| - B-LOCATION | |
| - I-LOCATION | |
| - B-TIMEDATE | |
| - I-TIMEDATE | |
| - B-PRODUCT | |
| - I-PRODUCT | |
| - B-WORKOFART | |
| - I-WORKOFART | |
| - B-JOB | |
| - I-JOB | |
| - B-AMOUNT | |
| - I-AMOUNT | |