# Stable Diffusion Trainer

> Efficiently train custom Stable Diffusion models with flexible batch sizes, gradient checkpointing, and memory-optimized attention requiring 12-24 GB VRAM for high-quality 512×512 to 1024×1024 image outputs.

## Overview

- **Model ID**: `sd-trainer`
- **Category**: training
- **Provider**: modelslab
- **Status**: active
- **Screenshot**: `https://assets.modelslab.com/generations/f352b8d0-1613-415c-9215-07cb5829ddcb.webp`

## API Information

This model can be used via our HTTP API. See the API documentation and usage examples below.

### Endpoint

- **URL**: `https://modelslab.com/api/v6/trainer/train`
- **Method**: POST

### Parameters

- **`instance_prompt`** (required): Prompt for training the instance.
  - Type: textarea

- **`images`** (required): Dataset for training, either as a path string or a list of paths (Required)
  - Type: array

- **`hf_token`** (optional): Hugging Face token for accessing models.
  - Type: text

- **`hf_username`** (optional): Hugging Face username for model storage
  - Type: text

- **`base_model_type`** (optional): Type of the base model. (Required)
  - Type: text

- **`seed`** (optional): Random seed for reproducibility. (Required)
  - Type: number

- **`training_steps`** (optional): 
  - Type: number

- **`resolution`** (optional): Resolution for the training images.
  - Type: number (range: 512-1024)

- **`batch_size`** (optional): Batch size for training.
  - Type: number (range: 1-128)

- **`mixed_precision`** (optional): Precision mode for training allowed (fp16, bf16”)
  - Type: select (options: fp16)

- **`rank`** (optional): Rank for LoRA
  - Type: number (range: 1-512)

- **`alpha`** (optional): Alpha for LoRA.
  - Type: number (range: 1-1024)

- **`learning_rate`** (optional): 
  - Type: number

- **`server_name`** (optional): 
  - Type: select (options: NVIDIA GeForce RTX 3090 ($0.46/hr))

- **`trainer_id`** (optional): 
  - Type: text

## Usage Examples

### cURL

```bash
curl --request POST \
  --url https://modelslab.com/api/v6/trainer/train \
  --header "Content-Type: application/json" \
  --data '{
    "key": "YOUR_API_KEY",
    "model_id": "sd-trainer",
    "instance_prompt": "thalman than",
    "base_model_type": "sdxl",
    "seed": "298329",
    "training_steps": "1500",
    "resolution": "1024",
    "batch_size": "4",
    "mixed_precision": "fp16",
    "rank": "4",
    "alpha": "2",
    "learning_rate": "0.0001",
    "server_name": "NVIDIA GeForce RTX 3090 ($0.46/hr)",
    "trainer_id": "sd-trainer"
  }'
```

### Python

```python
import requests

response = requests.post(
    "https://modelslab.com/api/v6/trainer/train",
    headers={
        "Content-Type": "application/json"
    },
    json={
        "key": "YOUR_API_KEY",
        "model_id": "sd-trainer",
        "instance_prompt": "thalman than",
        "base_model_type": "sdxl",
        "seed": "298329",
        "training_steps": "1500",
        "resolution": "1024",
        "batch_size": "4",
        "mixed_precision": "fp16",
        "rank": "4",
        "alpha": "2",
        "learning_rate": "0.0001",
        "server_name": "NVIDIA GeForce RTX 3090 ($0.46/hr)",
        "trainer_id": "sd-trainer"
    }
)

print(response.json())
```

### JavaScript

```javascript
fetch("https://modelslab.com/api/v6/trainer/train", {
  method: "POST",
  headers: {
    "Content-Type": "application/json"
  },
  body: JSON.stringify({
    "key": "YOUR_API_KEY",
    "model_id": "sd-trainer",
    "instance_prompt": "thalman than",
    "base_model_type": "sdxl",
    "seed": "298329",
    "training_steps": "1500",
    "resolution": "1024",
    "batch_size": "4",
    "mixed_precision": "fp16",
    "rank": "4",
    "alpha": "2",
    "learning_rate": "0.0001",
    "server_name": "NVIDIA GeForce RTX 3090 ($0.46/hr)",
    "trainer_id": "sd-trainer"
  })
})
.then(response => response.json())
.then(data => console.log(data));
```

## Links

- [Model Playground](https://modelslab.com/models/sd-trainer/sd-trainer)
- [API Documentation](https://docs.modelslab.com)
- [ModelsLab Platform](https://modelslab.com)