jbochi/coedit-base

1---2license: apache-2.03base_model: google/flan-t5-base4datasets:5- grammarly/coedit6tags:7- generated_from_trainer8- text-generation-inference9metrics:10- rouge11model-index:12- name: coedit-base13  results: []14language:15- en16widget:17- text: >-18    Fix the grammar: When I grow up, I start to understand what he said is quite19    right.20  example_title: Fluency21- text: >-22    Make this text coherent: Their flight is weak. They run quickly through the23    tree canopy.24  example_title: Coherence25- text: >-26    Rewrite to make this easier to understand: A storm surge is what forecasters27    consider a hurricane's most treacherous aspect.28  example_title: Simplification29- text: 'Paraphrase this: Do you know where I was born?'30  example_title: Paraphrase31- text: 'Write this more formally: omg i love that song im listening to it right now'32  example_title: Formalize33- text: 'Write in a more neutral way: The authors'' exposé on nutrition studies.'34  example_title: Neutralize35---36 37<!-- This model card has been generated automatically according to the information the Trainer had access to. You38should probably proofread and complete it, then remove this comment. -->39 40# coedit-base41 42This model is a fine-tuned version of [google/flan-t5-base](https://huggingface.co/google/flan-t5-base) on the [CoEdIT dataset](https://huggingface.co/datasets/grammarly/coedit).43 44It achieves the following results on the evaluation set:45- Loss: 0.597846- Rouge1: 60.593147- Rouge2: 48.016548- Rougel: 57.899749- Rougelsum: 57.933550- Gen Len: 16.672951 52## Model description53 54More information needed55 56## Intended uses & limitations57 58More information needed59 60## Training and evaluation data61 62More information needed63 64## Training procedure65 66### Training hyperparameters67 68The following hyperparameters were used during training:69- learning_rate: 1e-0570- train_batch_size: 1071- eval_batch_size: 1072- seed: 4273- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-0874- lr_scheduler_type: linear75- num_epochs: 576 77### Training results78 79| Training Loss | Epoch | Step  | Validation Loss | Rouge1  | Rouge2  | Rougel  | Rougelsum | Gen Len |80|:-------------:|:-----:|:-----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:-------:|81| 0.7478        | 1.0   | 6908  | 0.6452          | 59.7569 | 46.3099 | 56.4301 | 56.4464   | 16.6268 |82| 0.7127        | 2.0   | 13816 | 0.6086          | 60.2082 | 47.27   | 57.2356 | 57.2531   | 16.6513 |83| 0.7136        | 3.0   | 20724 | 0.6059          | 60.3747 | 47.6257 | 57.595  | 57.6184   | 16.6349 |84| 0.7038        | 4.0   | 27632 | 0.5999          | 60.5075 | 47.7856 | 57.7316 | 57.7698   | 16.6735 |85| 0.6911        | 5.0   | 34540 | 0.5978          | 60.5931 | 48.0165 | 57.8997 | 57.9335   | 16.6729 |86 87 88### Framework versions89 90- Transformers 4.35.291- Pytorch 2.1.0+cu11892- Datasets 2.14.793- Tokenizers 0.15.0