|
1 | 1 | # Persona Modules |
2 | 2 |
|
3 | | -Persona modules are things like: |
| 3 | +Persona modules define the character and behavior of AI agents. These modules have gained significant traction with the rise of AI agents, enabling more natural and specialized agent interactions. |
4 | 4 |
|
| 5 | +Available collections on HuggingFace: |
5 | 6 | - [Social personas](https://huggingface.co/datasets/NapthaAI/social_agents_personas) |
6 | 7 | - [Market personas](https://huggingface.co/datasets/NapthaAI/market_agents_personas) |
7 | 8 |
|
8 | | -Personas can be synthetically generated or generated from real personal data, such as that exported from social or finance apps. |
| 9 | +Personas can be either synthetically generated using LLMs for specific use cases, or generated from real personal data (like social media and finance app exports) to preserve authentic behaviors. |
9 | 10 |
|
10 | 11 | ## Generate a Persona from your X Data |
11 | 12 |
|
12 | | -First, download an archive of your X data [here](https://help.x.com/en/managing-your-account/how-to-download-your-x-archive). |
13 | | - |
14 | | -Then, run tweets2character directly from your command line: |
| 13 | +1. Download an archive of your X data [here](https://twitter.com/settings/download_your_data). |
15 | 14 |
|
| 15 | +2. Run tweets2character directly from your command line: |
| 16 | +```bash |
16 | 17 | npx tweets2character |
17 | | -NOTE: You need an API key to use Claude or OpenAI. |
18 | | - |
19 | | -If everything is correct, you'll see a loading bar as the script processes your data and generates a character file. |
20 | | - |
21 | | -It will be output as character.json in the directory where you run npx tweets2character. |
22 | | - |
23 | | -Finally, submit a pull request to add your persona to this dataset on HuggingFace. |
24 | | - |
| 18 | +``` |
| 19 | +:::note |
| 20 | +You need an API key to use Claude or OpenAI. |
| 21 | +::: |
25 | 22 |
|
| 23 | +3. If everything is correct, you'll see a loading bar as the script processes your data and generates a character file. |
| 24 | +:::info |
| 25 | +It will be output as `character.json` in the directory where you run `npx tweets2character`. |
| 26 | +::: |
26 | 27 |
|
27 | | -You can follow and request to join the Naptha organization on Hugging Face [here](https://huggingface.co/NapthaAI) (this could take 24 hours or more) |
| 28 | +4. Finally, submit a pull request to add your persona to this dataset on HuggingFace. |
28 | 29 |
|
| 30 | +5. You can follow and request to join the Naptha organization on Hugging Face [here](https://huggingface.co/NapthaAI) (this could take 24 hours or more) |
29 | 31 |
|
30 | 32 | ```bash |
31 | 33 | unzip twitter-2024-08-19-7604d23503c4857295c24edc2a13ec3d6d972639076041cc58eedefa8b439e62.zip twitter-2024-08-19-7604d23503c4857295c24edc2a13ec3d6d972639076041cc58eedefa8b439e62 |
32 | 34 | ``` |
33 | | - |
34 | 35 | ```bash |
35 | 36 | naptha write_storage -i /Users/arshath/play/tweets_to_character/twitter-2024-11-14-ebb9578b384ebab9a263b7621eb86794462f7f5fa47d51d2e33a7607ed0d8f70 |
36 | 37 | ``` |
37 | | - |
38 | 38 | ```bash |
39 | 39 | # usage: naptha run <agent_name> <agent args> |
40 | 40 | naptha run agent:tweets_to_character -p "input_dir=twitter-2024-08-19-7604d23503c4857295c24edc2a13ec3d6d972639076041cc58eedefa8b439e62" |
41 | 41 | ``` |
42 | 42 |
|
| 43 | +## Need Help? |
| 44 | +- Join our [Discord](https://discord.gg/naptha) |
| 45 | +- Submit issues on [GitHub](https://github.com/NapthaAI) |
0 commit comments