Skip to main content

Week 3

July 07/14/23​

Plan for the openai exploration​

Models to explore

'gpt-4', 'gpt-4-0613', 'gpt-3.5-turbo', 'gpt-3.5-turbo-0613', 'text-davinci-003',

July 07/13/23​

Need to add updates

July 07/12/23​

  1. Docs
  2. Colab notebooks
    1. The skip_exec flag is added to all the notebooks, that needs to be taken out at some point.
  3. Vannary
    1. Send out an email to Vannary
  4. AI3
    1. Push the code
    2. Drop the abbreviations table for now
    3. Add the lookup function to return the closest match based on distance
    4. Try experiments by revising the prompts
    5. Start the routine to ask Bard questions
  5. Pipe breaks app
    1. Look at its current state
    2. Create issues and note down possible places of refactor

Email for Vannary​

Hey Vannary,

I had question about the DNV download that I had done a couple of weeks ago.

Bard experiment​

provide your insight in 100 words or less​

The most important takeaway from this data is that the officer's cognitive level is yellow, which is a warning sign that they may be experiencing cognitive fatigue. This is supported by the fact that the officer's EEG beta value is low, which indicates that they are not as alert as they should be. Additionally, the officer's eye track microsaccade magnitude is high, which suggests that they are having difficulty focusing. As a result, the cognitive monitor has issued an alert, and the alert interface bodycam has been automatically turned on. This will allow the officer's supervisor to monitor their condition and intervene if necessary.

In short, this data suggests that the officer is experiencing cognitive fatigue and may not be able to perform their duties safely. The alert system has been activated to help protect the officer and the public.

Here are some additional details from the data that support the conclusion that the officer is experiencing cognitive fatigue:

  • The sensor electrodermal activity value is high, which indicates that the officer is stressed.
  • The sensor ambient light value is high, which could be distracting the officer.
  • The sensor pupil diameter value is low, which suggests that the officer's pupils are constricted, which can be a sign of fatigue.
  • The sensor eye track avg fix duration value is low, which indicates that the officer is not able to focus their eyes for very long.

Overall, the data suggests that the officer is experiencing cognitive fatigue and may not be able to perform their duties safely. The alert system has been activated to help protect the officer and the public.

75 words or less​

Sure, here is my insight in 75 words or less:

The most important takeaway from this data is that the officer is experiencing cognitive fatigue, as indicated by their low EEG beta value and high eye track microsaccade magnitude. This has triggered an alert, and the bodycam has been automatically turned on to allow the officer's supervisor to monitor their condition.

Here are the key details from the data that support this conclusion:

  • EEG beta value: 6 uv (low)
  • Eye track microsaccade magnitude: 1 deg (high)
  • Alert level: Yellow
  • Alert interface bodycam: Auto_On_Orange

Overall, the data suggests that the officer is experiencing cognitive fatigue and may not be able to perform their duties safely. The alert system has been activated to help protect the officer and the public.

50 words or less​

Sure, here is my insight in 50 words or less:

The officer is experiencing cognitive fatigue, as indicated by their low EEG beta value and high eye track microsaccade magnitude. This has triggered an alert, and the bodycam has been automatically turned on.

Here are the key details from the data that support this conclusion:

  • EEG beta value: 6 uv (low)
  • Eye track microsaccade magnitude: 1 deg (high)
  • Alert level: Yellow
  • Alert interface bodycam: Auto_On_Orange

Overall, the data suggests that the officer is experiencing cognitive fatigue and may not be able to perform their duties safely. The alert system has been activated to help protect the officer and the public.

July 07/11/23​

Colab notebooks migration plan​

  1. Download all the notebooks in Google Colab with the help of rclone
  2. Categorize the notebooks into their appropriate 7 categories
    1. Have the list finalized with Sudhir
  3. Create a new nbdev style repo called experiments
    1. Categorize the notebooks related to experiments in this repo
    2. Discuss with Sudhir
  4. Move notebooks related to hackathons in the hackathons repo
  5. Migrate the rest of the notebooks to gqc-utility-notebooks
  6. Once everything is push to their apporpriate repos, remove all the notebooks from Google Colab
info

A new repo for notebooks-from-colab was created. All the 160 notebooks will be pushed over there. At a later time, slowly the notebooks will be re-organized and moved to other repositories.

July 07/10/23​

  1. Docs
    1. Finish geopandas minor walkthrough
    2. Colab notebooks organization

Organizing colab notebooks​

Most of these notebooks can be split into these 6 broad categories

  1. computer vision
  2. nlp
  3. time series forecasting
  4. machine learning (tabular data)
  5. hackathons
  6. experiments
    1. experiments related to other GQC projects
    2. random experiments
    3. hyperparametering tuning and experiment tracking
Last updated on by Deven