Instructions to use research-backup/mbart-large-cc25-itquad-qa with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use research-backup/mbart-large-cc25-itquad-qa with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="research-backup/mbart-large-cc25-itquad-qa")# Load model directly from transformers import AutoTokenizer, AutoModelForSeq2SeqLM tokenizer = AutoTokenizer.from_pretrained("research-backup/mbart-large-cc25-itquad-qa") model = AutoModelForSeq2SeqLM.from_pretrained("research-backup/mbart-large-cc25-itquad-qa") - Notebooks
- Google Colab
- Kaggle
- Local Apps
- vLLM
How to use research-backup/mbart-large-cc25-itquad-qa with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "research-backup/mbart-large-cc25-itquad-qa" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "research-backup/mbart-large-cc25-itquad-qa", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/research-backup/mbart-large-cc25-itquad-qa
- SGLang
How to use research-backup/mbart-large-cc25-itquad-qa with SGLang:
Install from pip and serve model
# Install SGLang from pip: pip install sglang # Start the SGLang server: python3 -m sglang.launch_server \ --model-path "research-backup/mbart-large-cc25-itquad-qa" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "research-backup/mbart-large-cc25-itquad-qa", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker images
docker run --gpus all \ --shm-size 32g \ -p 30000:30000 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ --env "HF_TOKEN=<secret>" \ --ipc=host \ lmsysorg/sglang:latest \ python3 -m sglang.launch_server \ --model-path "research-backup/mbart-large-cc25-itquad-qa" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "research-backup/mbart-large-cc25-itquad-qa", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use research-backup/mbart-large-cc25-itquad-qa with Docker Model Runner:
docker model run hf.co/research-backup/mbart-large-cc25-itquad-qa
YAML Metadata Warning:The pipeline tag "text2text-generation" is not in the official list: text-classification, token-classification, table-question-answering, question-answering, zero-shot-classification, translation, summarization, feature-extraction, text-generation, fill-mask, sentence-similarity, text-to-speech, text-to-audio, automatic-speech-recognition, audio-to-audio, audio-classification, audio-text-to-text, voice-activity-detection, depth-estimation, image-classification, object-detection, image-segmentation, text-to-image, image-to-text, image-to-image, image-to-video, unconditional-image-generation, video-classification, reinforcement-learning, robotics, tabular-classification, tabular-regression, tabular-to-text, table-to-text, multiple-choice, text-ranking, text-retrieval, time-series-forecasting, text-to-video, image-text-to-text, image-text-to-image, image-text-to-video, visual-question-answering, document-question-answering, zero-shot-image-classification, graph-ml, mask-generation, zero-shot-object-detection, text-to-3d, image-to-3d, image-feature-extraction, video-text-to-text, keypoint-detection, visual-document-retrieval, any-to-any, video-to-video, other
Model Card of lmqg/mbart-large-cc25-itquad-qa
This model is fine-tuned version of facebook/mbart-large-cc25 for question answering task on the lmqg/qg_itquad (dataset_name: default) via lmqg.
Overview
- Language model: facebook/mbart-large-cc25
- Language: it
- Training data: lmqg/qg_itquad (default)
- Online Demo: https://autoqg.net/
- Repository: https://github.com/asahi417/lm-question-generation
- Paper: https://arxiv.org/abs/2210.03992
Usage
- With
lmqg
from lmqg import TransformersQG
# initialize model
model = TransformersQG(language="it", model="lmqg/mbart-large-cc25-itquad-qa")
# model prediction
answers = model.answer_q(list_question="Quale batterio ha il nome del paese che colpisce di più nel suo nome?", list_context=" Il complesso M. tubercolosi (MTBC) comprende altri quattro micobatteri causa di tubercolosi: M. bovis, M. africanum, M. canetti e M. microti. M. africanum non è molto diffuso, ma è una causa significativa di tubercolosi in alcune parti dell' Africa. M. bovis era una volta una causa comune della tubercolosi, ma l' introduzione del latte pastorizzato ha quasi completamente eliminato questo problema di salute pubblica nei paesi sviluppati. M. canetti è raro e sembra essere limitato al Corno d' Africa, anche se alcuni casi sono stati osservati negli emigranti africani. M. microti è anche raro ed è visto quasi solo in persone immunodeficienti, anche se la sua prevalenza può essere significativamente sottovalutata.")
- With
transformers
from transformers import pipeline
pipe = pipeline("text2text-generation", "lmqg/mbart-large-cc25-itquad-qa")
output = pipe("question: Quale batterio ha il nome del paese che colpisce di più nel suo nome?, context: Il complesso M. tubercolosi (MTBC) comprende altri quattro micobatteri causa di tubercolosi: M. bovis, M. africanum, M. canetti e M. microti. M. africanum non è molto diffuso, ma è una causa significativa di tubercolosi in alcune parti dell' Africa. M. bovis era una volta una causa comune della tubercolosi, ma l' introduzione del latte pastorizzato ha quasi completamente eliminato questo problema di salute pubblica nei paesi sviluppati. M. canetti è raro e sembra essere limitato al Corno d' Africa, anche se alcuni casi sono stati osservati negli emigranti africani. M. microti è anche raro ed è visto quasi solo in persone immunodeficienti, anche se la sua prevalenza può essere significativamente sottovalutata.")
Evaluation
- Metric (Question Answering): raw metric file
| Score | Type | Dataset | |
|---|---|---|---|
| AnswerExactMatch | 50.01 | default | lmqg/qg_itquad |
| AnswerF1Score | 64.73 | default | lmqg/qg_itquad |
| BERTScore | 93.12 | default | lmqg/qg_itquad |
| Bleu_1 | 32.51 | default | lmqg/qg_itquad |
| Bleu_2 | 26.71 | default | lmqg/qg_itquad |
| Bleu_3 | 22.92 | default | lmqg/qg_itquad |
| Bleu_4 | 19.64 | default | lmqg/qg_itquad |
| METEOR | 33.6 | default | lmqg/qg_itquad |
| MoverScore | 80.49 | default | lmqg/qg_itquad |
| ROUGE_L | 37.59 | default | lmqg/qg_itquad |
Training hyperparameters
The following hyperparameters were used during fine-tuning:
- dataset_path: lmqg/qg_itquad
- dataset_name: default
- input_types: ['paragraph_question']
- output_types: ['answer']
- prefix_types: None
- model: facebook/mbart-large-cc25
- max_length: 512
- max_length_output: 32
- epoch: 16
- batch: 4
- lr: 0.0001
- fp16: False
- random_seed: 1
- gradient_accumulation_steps: 16
- label_smoothing: 0.15
The full configuration can be found at fine-tuning config file.
Citation
@inproceedings{ushio-etal-2022-generative,
title = "{G}enerative {L}anguage {M}odels for {P}aragraph-{L}evel {Q}uestion {G}eneration",
author = "Ushio, Asahi and
Alva-Manchego, Fernando and
Camacho-Collados, Jose",
booktitle = "Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing",
month = dec,
year = "2022",
address = "Abu Dhabi, U.A.E.",
publisher = "Association for Computational Linguistics",
}
- Downloads last month
- 9
Dataset used to train research-backup/mbart-large-cc25-itquad-qa
Paper for research-backup/mbart-large-cc25-itquad-qa
Evaluation results
- BLEU4 (Question Answering) on lmqg/qg_itquadself-reported19.640
- ROUGE-L (Question Answering) on lmqg/qg_itquadself-reported37.590
- METEOR (Question Answering) on lmqg/qg_itquadself-reported33.600
- BERTScore (Question Answering) on lmqg/qg_itquadself-reported93.120
- MoverScore (Question Answering) on lmqg/qg_itquadself-reported80.490
- AnswerF1Score (Question Answering) on lmqg/qg_itquadself-reported64.730
- AnswerExactMatch (Question Answering) on lmqg/qg_itquadself-reported50.010