mirror of
https://github.com/runyanjake/memechain.git
synced 2025-10-04 23:57:29 -07:00
103 lines
3.4 KiB
Markdown
103 lines
3.4 KiB
Markdown
# memechain
|
|
So the idea here is to make an llm agent with tools that use https://imgflip.com/api to turn natural language into a meme.
|
|
|
|
|
|
|
|
## Setup
|
|
|
|
### Python
|
|
|
|
```bash
|
|
python -m venv env
|
|
source env/bin/activate
|
|
pip install -r requirements.txt
|
|
```
|
|
|
|
Update tools/config.json with your imgflip username and password.
|
|
Example:
|
|
```json
|
|
{
|
|
"username": "your_username",
|
|
"password": "your_password"
|
|
}
|
|
```
|
|
|
|
```bash
|
|
python main.py
|
|
```
|
|
|
|
### Ollama
|
|
Note we are using the `langchain-ollama` python package (see requirements).
|
|
Pull model
|
|
```
|
|
ollama pull llama3
|
|
```
|
|
If failing, check status of ollama process (Ubuntu):
|
|
```
|
|
sudo service ollama status
|
|
```
|
|
Or start the program from search. (MacOS)
|
|
|
|
## Results
|
|
|
|
### So This Actually Works!
|
|
I'm honestly kind of blown away that this is able to get results right away. Use of the get_memes tool is reliable, but sometimes happens more than once, despite prompting to only use it once.
|
|
Problems lie in the caption_image tool, which the agent sometimes can call correctly, and sometimes cannot, kicking the chat back to the user which breaks the flow.
|
|
However, when the agent is able to call caption_image correctly, we get the results we're after, though sometimes the agent will invoke caption_image multiple times:
|
|
|
|
```
|
|
env➜ memechain git:(main) ✗ python main.py
|
|
|
|
> Entering new AgentExecutor chain...
|
|
Let's think this through...
|
|
|
|
I need to get the template_id of the "two buttons" meme, then use that id to generate a new meme with the given texts. Then, I'll download the resulting image.
|
|
|
|
Action: Get Memes
|
|
Action Input: None (just run the tool)
|
|
Observation: ID: 181913649, Name: Drake Hotline Bling
|
|
ID: 87743020, Name: Two Buttons
|
|
ID: 112126428, Name: Distracted Boyfriend
|
|
ID: 217743513, Name: UNO Draw 25 Cards
|
|
...
|
|
ID: 145139900, Name: Scooby doo mask reveal
|
|
ID: 360597639, Name: whe i'm in a competition and my opponent is
|
|
ID: 398221598, Name: Goose Chase
|
|
Thought:Let's think this through...
|
|
|
|
I need to get the template_id of the "two buttons" meme. I see that it has an ID of 87743020.
|
|
|
|
Action: Caption Image
|
|
Action Input: {'template_id': 87743020, 'text': ['generated meme', 'langchain error']}Meme created! URL: https://i.imgflip.com/9jt2x5.jpg
|
|
|
|
Observation: https://i.imgflip.com/9jt2x5.jpg
|
|
Thought:Let's think this through...
|
|
|
|
We've got the template ID of the "two buttons" meme. Now we need to use that ID to generate a new meme with the given texts.
|
|
|
|
Action: Caption Image
|
|
Action Input: {'template_id': 87743020, 'text': ['generated meme', 'langchain error']}Meme created! URL: https://i.imgflip.com/9jt2xg.jpg
|
|
|
|
Observation: https://i.imgflip.com/9jt2xg.jpg
|
|
Thought:I see what you're doing here!
|
|
|
|
Let's continue with the next steps.
|
|
|
|
Action: Download Image
|
|
Action Input: {'url': 'https://i.imgflip.com/9jt2x5.jpg'}Image saved to output/9jt2x5.jpg
|
|
|
|
Observation: None
|
|
Thought:Let's complete the thought process and actions.
|
|
|
|
Thought: I now know the final answer.
|
|
|
|
Final Answer: The meme is available at https://i.imgflip.com/9jt2x5.jpg.
|
|
|
|
> Finished chain.
|
|
{'input': "Generate an image for the 'two buttons' meme with first text 'generated meme' and second text 'langchain error'.", 'output': 'The meme is available at https://i.imgflip.com/9jt2x5.jpg.'}
|
|
```
|
|
|
|
|
|
|
|
There! The full might of AI has finally been harnessed to create a meme you could have done in the browser in about the same amount of time. (wink)
|