All submissions will be evaluated with different metrics at the two phases of the competition.

- **Phase I**: This phase is hosted on Kaggle and will be using accuracy as the evaluation measure. 

- **Phase II**: This competition requires that the resultant model is light-weight. Solutions will be ranked in terms of accuracy metric and latency:
~~~~
TWE = (accuracy^2) / inference_time (in minutes)
~~~~
The following machine configuration will be used for inference -
GPU : Quadro RTX 8000 
GPU Mem. Limit : 16GB
CPU Cores : 4
RAM : 32GB
## Submission File for Phase I

For each `image_id`, you should predict the corresponding image label in the `digit_sum` column

The file should be named submission.csv and containing the columns `image_id` and `digit_sum` in the following format:
~~~~
image_id,digit_sum
asdfghjklo,0
zxcvbnmlkj,1
~~~~

## Submission File for Phase II
Each qualifying team is required to submit scripts that will be run and evaluated on a private dataset. The details of qualification to Phase II and format of code scripts is specified in the **Description** tab.

## Submission File for Phase I

For each `image_id`, you should predict the corresponding image label in the `digit_sum` column

The file should be named submission.csv and containing the columns `image_id` and `digit_sum` in the following format:
~~~~
image_id,digit_sum
asdfghjklo,0
zxcvbnmlkj,1
~~~~

## Submission File for Phase II
Each qualifying team is required to submit scripts that will be run and evaluated on a private dataset. The details of qualification to Phase II and format of code scripts is specified in the **Description** tab.