# Final Project
###### tags: `project`
::: danger
**2020/06/24 FINAL REPORT**
An e3 assignment is now available for submitting the final report.
Due date: 2020/7/7 22:00. There are no extensions (Grade 0 after due date.)
* Submit a PDF file following the [ACL 2020 LaTeX Template](Templatehttps://www.overleaf.com/latex/templates/acl-2020-proceedings-template/zsrkcwjptpcd).
* To see samples of previous technical reports, check the reports submitted for the EmotionX 2019 competition here: https://sites.google.com/view/emotionx2019/reports
* Note that the EmotionX 2019 website contains 1 overview paper and 7 technical reports. Your report should follow a similar structure to the technical reports.
* We will upload an an overview paper for EmotionGIF 2020 on around July 5~7.
Download the ACL 2020 template here:
:::
::: info
**2020/06/08 IMPORTANT**
Due to the impact of this year's unique circumstances on the ACL2020 conference and workshops, we will be running a second round ("Round 2") of the EmotionGIF 2020 competition.
If you haven't participated in the original competition ("Round 1"), you can use this opportunity and have your final project participate in a live competition.
The datasets, metrics, instructions etc. are identical to the first round. The only thing that is new are the dates, which you can check on the [EmotionGIF website](https://sites.google.com/view/emotiongif-2020/important-dates).
Here are the link to the new evaluation websites:
[Round 2, Main Track](https://competitions.codalab.org/competitions/25174)
[Round 2, Auxiliary Track](https://competitions.codalab.org/competitions/25175)
:::
::: warning
If you have a Twitter account, you can follow [EmotionGIF2020](https://twitter.com/EmotionGIF2020).
:::
In the final project you wil participate in the EmotionGIF 2020 challenge.
The challenge is to predict the GIF response to text.
We suggest you complete the following steps as soon as possible to familiarize yourself with the competition.
Here are the first steps:
* [Read about the competition](https://sites.google.com/view/emotiongif-2020/)
* [Submit the form](https://sites.google.com/view/emotiongif-2020/participate)
* You will get an automatic email with a link to the Codalab evaluation website
* Register at the Codalab evaluation website
* Create a Codalab team in order to participate (teams can have between 1~3 students)
Here is a famous example of a GIF. The dataset contains many others :)

## Discussion
:::success
This forum is used for asking question and sharing information.
Students are encouraged to share their knowledge or solutions with other students.
In addition, the forum will be checked by the TAs every 12~24 hours.
:::
> I found I have some error messages when I submitting my dev.zip, but I actually couldn't figure it out why happens. I compare with my previous successful output I didn't find some strange format. Log is `WARNING: Your kernel does not support swap limit capabilities or the cgroup is not mounted. Memory limited without swap.` But after few hours, it will become success and get scores.
>> You can safely ignore the WARNING. Usually a score will not appear if you have a problem with your submission, so look carefully at the output. One more point: after you successfully submit, click on the UPDATE button. After that you will need to REFRESH the browser page to get the score. Hope that helps! [name=Boaz][color=#ff00ff]
>>
> I found that all my descriptions submitted yesterday are all missed, but records submitted before still exist. Where can I find this? Thanks.
> > Sometimes the Codalab platform acts weird. (Maybe because Microsoft was involved...?) I suggest keeping a copy of your submissions, and simply submit again.[name=Boaz][color=#ff00ff]
> There is an aggrement says `participants are not allowed to use external data or custom-trained models in their solutions. Participants are allowed to use publicly available pretrained models and embeddings, such as BERT and GloVe.` What do `custom-trained models` and `publicly available pretrained models` mean? If someone shared his fine-tuned model publicly, is this model publicly available pretrained model? Another question is teacher said earlier today we can use external data to get better performance, is this allowed in this competition?
>>
>> In the [Main Track](https://competitions.codalab.org/competitions/24761), you should use only standard, publicly available pre-trained models and word embeddings (BERT, GloVe, and similar), and no external data. The purpose is to create an even playing field for the competition.
>> There is also an [Auxilary track](https://competitions.codalab.org/competitions/24820), where you can submit solutions with external data, custom pre-trained models, etc.[name=Boaz][color=#ff00ff]
> Now it is the evaluation phase for the final competition. I noticed that the total submission is 10. Is that means we can only submit 10 times in this phase? If so, can we still submit the dev.json in unlimited times for practice of our model?
>> These dates do not apply to NCTU students. More details about the final project will be available later. For now, make sure to submit the assignment on the [Lab 5](https://competitions.codalab.org/competitions/25115?secret_key=e3d89637-74fe-41e4-acb5-c734d91edebe) website. Also, you can use the website to submit as many predictions as you like if you want to test any other model. [name=Boaz][color=#ff00ff]
>>> I remember teacher said our final is to try this competition. I thought this mean we should follow competition schedule instead of starting after formal competition?
>>>> The competition data is exactly the same, but you don't need to follow the official dates. You can use the [Lab 5](https://competitions.codalab.org/competitions/25115?secret_key=e3d89637-74fe-41e4-acb5-c734d91edebe) website to practice.[name=Boaz][color=#ff00ff]
>>>>> Actually what I confuse is whether we must follow official dates **mentioned by teacher**? Also, if I can still submit dev data on Lab 5, won't it be unfair of official competition? Thanks.
> Do we need to upload our code after the evaluation phase(UTC. 6.8 8:00)? If we do, where to submit?
>> Please see the answer(s) to the previous question.
> If we submit at the 'post-evaluation', is our submission in evaluation phase still effective?
>> Please see the announcement at the top of the page. You can submit your predictions at the Round 2 Main Track now.[name=Boaz][color=#ff00ff]