Summarization
Transformers
Safetensors
Estonian
mbart
text2text-generation
mbart50
low-resource
research
rahvusarhiiv
Eval Results (legacy)
Instructions to use Rahvusarhiiv/et_summariser with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Rahvusarhiiv/et_summariser with Transformers:
# Use a pipeline as a high-level helper # Warning: Pipeline type "summarization" is no longer supported in transformers v5. # You must load the model directly (see below) or downgrade to v4.x with: # 'pip install "transformers<5.0.0' from transformers import pipeline pipe = pipeline("summarization", model="Rahvusarhiiv/et_summariser")# Load model directly from transformers import AutoTokenizer, AutoModelForSeq2SeqLM tokenizer = AutoTokenizer.from_pretrained("Rahvusarhiiv/et_summariser") model = AutoModelForSeq2SeqLM.from_pretrained("Rahvusarhiiv/et_summariser", device_map="auto") - Notebooks
- Google Colab
- Kaggle
| { | |
| "repo_id": "Rahvusarhiiv/et_summariser", | |
| "repo_name": "et_summariser", | |
| "language": "et", | |
| "mbart_lang_code": "et_EE", | |
| "base_model": "facebook/mbart-large-50-many-to-many-mmt", | |
| "local_model_dir": "outputs/dedicated_summarization_models/et_mbart50_grounded_v1/final_model", | |
| "training_metadata": "outputs/dedicated_summarization_models/et_mbart50_grounded_v1/training_metadata.json", | |
| "training_report": "outputs/summarization_training_data_grounded_clean_all_full_lv_v1/split_report.md", | |
| "release_stance": "experimental; no SOTA claim", | |
| "eval_rows": [ | |
| { | |
| "label": "Wiki4 strict v2 LLM judge", | |
| "dataset": "wiki4_et", | |
| "examples": "100", | |
| "factuality": 3.01, | |
| "coverage": 3.33, | |
| "focus": 4.27, | |
| "language_score": 4.3, | |
| "overall": 2.73, | |
| "kind": "ollama_judge" | |
| }, | |
| { | |
| "label": "Wiki4 broad LLM judge", | |
| "dataset": "wiki4_et", | |
| "examples": "200", | |
| "factuality": 3.02, | |
| "coverage": 2.42, | |
| "focus": 3.715, | |
| "language_score": 4.185, | |
| "overall": 2.33, | |
| "kind": "ollama_judge" | |
| }, | |
| { | |
| "label": "Wiki4 final automatic", | |
| "dataset": "wiki4_et", | |
| "examples": 100, | |
| "rougeL_f": 0.29355689324997836, | |
| "rouge1_f": 0.3513885249581917, | |
| "rouge2_f": 0.20140574646547016, | |
| "reference_content_token_coverage": 0.38743965026667565, | |
| "prediction_source_content_token_precision": 0.7509638190909874, | |
| "prediction_words": 62.39, | |
| "repeated_4gram_rate": 0.0, | |
| "eu_legalese_template_flag": 0.0, | |
| "hit_max_new_tokens": 0.0, | |
| "kind": "automatic" | |
| }, | |
| { | |
| "label": "GDELT final automatic", | |
| "dataset": "gdelt_et", | |
| "examples": 100, | |
| "rougeL_f": 0.1692342603250661, | |
| "rouge1_f": 0.21227974458081833, | |
| "rouge2_f": 0.07460760477756037, | |
| "reference_content_token_coverage": 0.2569681719805297, | |
| "prediction_source_content_token_precision": 0.8053276209384402, | |
| "prediction_words": 47.48, | |
| "repeated_4gram_rate": 0.0, | |
| "eu_legalese_template_flag": 0.0, | |
| "hit_max_new_tokens": 0.0, | |
| "kind": "automatic" | |
| } | |
| ] | |
| } | |