File size: 1,993 Bytes
871cdf4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
4deef57
5bc0f78
871cdf4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
---
language: en
license: mit
base_model: mlx-community/Qwen2.5-1.5B-Instruct-4bit
tags:
- mlx
- intent-detection
- fine-tuned
- knowledge-management
- ios
library_name: mlx
---

# knowledgebase-intent-llm

Fine-tuned model for intent detection in a knowledge management iOS app.

## Model Details

- **Base Model**: mlx-community/Qwen2.5-1.5B-Instruct-4bit
- **Model Type**: intent_detection
- **Format**: complete_merged
- **Framework**: MLX
- **Training Examples**: 5000
- **Training Iterations**: 100

## Usage

```python
from mlx_lm import load, generate

# Load the model
model, tokenizer = load("hebertgo/knowledgebase-intent-llm")

# Generate intent classification
prompt = '''You are a helpful AI assistant for a knowledge-management app on an iPhone. Analyze the user's request and respond with JSON in this format:
{
  "action": "Search|Create|Clarify|Conversation",
  "response": "User-friendly response message",
  "contentType": "videos|bookmarks|todos",
  "topic": "extracted topic or null"
}

User query: find videos about python'''

response = generate(model, tokenizer, prompt=prompt, max_tokens=256)
print(response)
```

## iOS Integration

This model is designed for use in iOS apps with MLX Swift:

```swift
let config = ModelConfiguration(
    id: "hebertgo/knowledgebase-intent-llm",
    defaultPrompt: ""
)

let model = try await LLMModelFactory.shared.loadContainer(
    configuration: config
)
```

## Training Details

- **Fine-tuning Method**: LoRA with model fusion
- **Export Date**: 2025-06-24T17:12:38.111419
- **Fusion Completed**: True

## Expected Outputs

The model generates JSON responses with these action types:
- **Search**: Find existing content (videos, bookmarks, todos)
- **Create**: Add new content 
- **Clarify**: Request more information
- **Conversation**: General chat responses

Content types supported:
- videos
- bookmarks  
- todos

## Performance

Optimized for Apple Silicon devices with MLX framework for efficient on-device inference.