---
license: apache-2.0
tags:
- generated_from_trainer
metrics:
- rouge
model-index:
- name: t5-base-DreamBank-Generation-Act-Char
  results: []
language:
- en
inference:
  parameters:
    max_length: 128
widget:
- text: "I was skating on the outdoor ice pond that used to be across the street from my house. I was not alone, but I did not recognize any of the other people who were skating around. I went through my whole repertoire of jumps, spires, and steps-some of which I can do and some of which I'm not yet sure of. They were all executed flawlessly-some I repeated, some I did only once. I seemed to know that if I went into competition, I would be sure of coming in third because there were only three contestants. Up to that time I hadn't considered it because I hadn't thought I was good enough, but now since everything was going so well, I decided to enter."
  example_title: "Dream"
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# t5-base-DreamBank-Generation-Act-Char

This model is a fine-tuned version of [DReAMy-lib/t5-base-DreamBank-Generation-NER-Char](https://huggingface.co/DReAMy-lib/t5-base-DreamBank-Generation-NER-Char) on the DreamBank dataset.
The uploaded model contains the weights of the best-performing model (see table below), tune to annotate a given
dream report according to [Hall and Van de Castle the Activity feature](https://dreams.ucsc.edu/Coding/activities.html)

## Training procedure

The model is trained end-to-end using a text2text solution to annotate dream reports following the Activity feature 
from the Hall and Van de Castle scoring framework. Given a report, the model generates texts of the form 
`[(initialiser : activity type : receiver)]`. For those cases where `initialiser` and `receiver` are the same 
entity, the output will follow the `[(initialiser : alone : activity type)]` setting.

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 0.001
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum |
|:-------------:|:-----:|:----:|:---------------:|:------:|:------:|:------:|:---------:|
| No log        | 1.0   | 49   | 0.4061          | 0.3684 | 0.2537 | 0.3495 | 0.3484    |
| No log        | 2.0   | 98   | 0.3563          | 0.4151 | 0.3185 | 0.4043 | 0.4030    |
| No log        | 3.0   | 147  | 0.3005          | 0.4456 | 0.3588 | 0.4294 | 0.4281    |
| No log        | 4.0   | 196  | 0.2693          | 0.4743 | 0.3903 | 0.4586 | 0.4574    |
| No log        | 5.0   | 245  | 0.2627          | 0.4751 | 0.3939 | 0.4564 | 0.4549    |
| No log        | 6.0   | 294  | 0.2739          | 0.4744 | 0.3920 | 0.4612 | 0.4596    |
| No log        | 7.0   | 343  | 0.2733          | 0.4702 | 0.3940 | 0.4557 | 0.4549    |
| No log        | 8.0   | 392  | 0.2861          | 0.4739 | 0.3950 | 0.4614 | 0.4608    |
| No log        | 9.0   | 441  | 0.3115          | 0.4645 | 0.3868 | 0.4524 | 0.4517    |
| No log        | 10.0  | 490  | 0.3212          | 0.4655 | 0.3886 | 0.4524 | 0.4518    |


### Framework versions

- Transformers 4.25.1
- Pytorch 1.12.1
- Datasets 2.5.1
- Tokenizers 0.12.1