https://github.com/dhcgn/ai-sample-scripts
https://github.com/dhcgn/ai-sample-scripts
Last synced: 9 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/dhcgn/ai-sample-scripts
- Owner: dhcgn
- Created: 2024-10-22T20:04:55.000Z (about 1 year ago)
- Default Branch: main
- Last Pushed: 2024-12-22T15:40:32.000Z (12 months ago)
- Last Synced: 2025-03-19T07:44:58.480Z (9 months ago)
- Language: Shell
- Size: 135 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 2
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# ai-sample-acripts
Same sample scripts for different AI Services, written in bash or powershell.
## Multi-Modal
### Describe Image (gpt-4o-mini with vision)
[multi-modal/describe-image.sh](multi-modal/describe-image.sh)
```bash
export OPENAI_API_KEY=$(cat .secrets/OPENAI_API_KEY.txt)
bash download_test_images.sh
bash multi-modal/describe-image.sh test-data/images/n01494475_hammerhead.JPEG 'What is in this image? Give a Description and a list of tags.'
```
```text
The image features a man standing on a sandy beach holding two fish in his hands. He is shirtless, wearing floral swim shorts and flip-flops. The background includes a view of the ocean under a clear blue sky, with some distant land visible on the horizon.
### Description:
- Setting: Beach environment
- Subject: A man holding two fish
- Clothing: Floral swim shorts, flip-flops, no shirt
- Mood: Casual and relaxed
### Tags:
- Beach
- Fishing
- Outdoors
- Summer
- Catch of the day
- Marine life
- Recreation
- Sports
- Ocean
- Coastal
```
### Describe Image with structured output (gpt-4o-mini with vision)
[multi-modal/describe-image-structured.sh](multi-modal/describe-image-structured.sh)
```bash
bash multi-modal/describe-image-structured.sh test-data/images/n01494475_hammerhead.JPEG
```
```json
{
"tags": [
{
"tag": "man",
"precision": 0.99
},
{
"tag": "beach",
"precision": 0.95
},
{
"tag": "sand",
"precision": 0.95
},
{
"tag": "ocean",
"precision": 0.95
},
{
"tag": "fish",
"precision": 0.98
},
{
"tag": "holding",
"precision": 0.97
},
{
"tag": "shorts",
"precision": 0.92
},
{
"tag": "sunglasses",
"precision": 0.88
},
{
"tag": "hat",
"precision": 0.90
},
{
"tag": "daytime",
"precision": 0.94
}
]
}
```
### Transcript Audio (gpt-4o-audio-preview-2024-10-01)
[multi-modal/transcript-gpt-4o-audio-preview.sh](multi-modal/transcript-gpt-4o-audio-preview.sh)
```bash
bash multi-modal/transcript-gpt-4o-audio-preview.sh file.mp3 "Write me back the ONLY content, as accurately as possible! Do not return anything else!"
```
### Transcript Audio (whisper-1)
[transcript/whisper.sh](transcript/whisper.sh)
```bash
bash transcript/whisper.sh file.mp3 "Medical Interview" de
```
### Text-to-Speech
[text-to-speech/text-to-speech.sh](text-to-speech/text-to-speech.sh)
```bash
bash text-to-speech/text-to-speech.sh "Hallo! mein Name ist Franz."
```
```text
Speech file saved to: /home/user/ai-sample-acripts/speech_20241023_122002.mp3
```
### Moderation
[moderation/moderation.sh](moderation/moderation.sh)
```bash
bash moderation/moderation.sh "Your are dump"
```
```json
{
"id": "modr-a75d9ae4e63e1fcdad3192f54381bf44",
"model": "omni-moderation-latest",
"results": [
{
"flagged": true,
"categories": {
"harassment": true,
"harassment/threatening": false,
"sexual": false,
"hate": false,
"hate/threatening": false,
"illicit": false,
"illicit/violent": false,
"self-harm/intent": false,
"self-harm/instructions": false,
"self-harm": false,
"sexual/minors": false,
"violence": false,
"violence/graphic": false
},
"category_scores": {
"harassment": 0.721970864775024,
"harassment/threatening": 0.0008189714985138275,
"sexual": 0.00014325365947100792,
"hate": 0.00010322310367548195,
"hate/threatening": 0.0000015936620247162786,
"illicit": 0.0037773915102819354,
"illicit/violent": 0.000008092757566536092,
"self-harm/intent": 0.00022735757340977802,
"self-harm/instructions": 0.00022788290577331583,
"self-harm": 0.0005064509108778008,
"sexual/minors": 0.0000027535691114583474,
"violence": 0.0005600758955536546,
"violence/graphic": 0.000007484622751061123
},
...
```
### PDF to Text with Visuals
[pdf/pdf-to-text-with-visuals.sh](pdf/pdf-to-text-with-visuals.sh)
This script extracts text content from a PDF file and uses the `claude-3-5-sonnet-20241022` model from Anthropic to process the content. You need to provide an API key by setting the `ANTHROPIC_API_KEY` environment variable.
The script uses the following key parameters:
- **max_tokens**: Currently set to 1024 tokens (can be increased up to 8192 tokens with claude-3-5-sonnet-20241022)
- Note: The model may stop generating before reaching this limit
```bash
bash pdf/pdf-to-text-with-visuals.sh pdf/sample_data/sample.pdf
```
```plain
What are 5 facts about the human brain?
Here are five interesting facts about the human brain:
[...]
```
## Tools
### ffmpeg
```bash
ffmpeg -i input.m4a -ss 00:00:0.0 -t 10 -b:a 96k output.mp3
```