|
|
--- |
|
|
license: apache-2.0 |
|
|
base_model: Mitchins/t5-base-artgen-multi-instruct |
|
|
tags: |
|
|
- text2text-generation |
|
|
- prompt-enhancement |
|
|
- ai-art |
|
|
- openvino |
|
|
- t5 |
|
|
- art-generation |
|
|
- stable-diffusion |
|
|
- intel |
|
|
language: |
|
|
- en |
|
|
library_name: optimum-intel |
|
|
pipeline_tag: text-generation |
|
|
model-index: |
|
|
- name: t5-base-artgen-multi-instruct-OpenVINO |
|
|
results: [] |
|
|
datasets: |
|
|
- art-prompts |
|
|
widget: |
|
|
- text: "Enhance this prompt: robot in space" |
|
|
example_title: "Standard Enhancement" |
|
|
- text: "Enhance this prompt (no lora): beautiful landscape" |
|
|
example_title: "Clean Enhancement" |
|
|
- text: "Enhance this prompt (with lora): anime girl" |
|
|
example_title: "Technical Enhancement" |
|
|
- text: "Simplify this prompt: A stunning, highly detailed masterpiece" |
|
|
example_title: "Simplification" |
|
|
--- |
|
|
|
|
|
# T5 Base Art Generation Multi-Instruct OpenVINO |
|
|
|
|
|
OpenVINO version of [Mitchins/t5-base-artgen-multi-instruct](https://huggingface.co/Mitchins/t5-base-artgen-multi-instruct) for optimized Intel hardware inference. |
|
|
|
|
|
## Model Details |
|
|
|
|
|
- **Base Model**: T5-base (Google) |
|
|
- **Training Samples**: 297,282 |
|
|
- **Parameters**: 222M |
|
|
- **Format**: OpenVINO IR (FP32) |
|
|
- **Optimization**: Intel CPU/GPU/VPU optimized |
|
|
|
|
|
## Quad-Instruction Capabilities |
|
|
|
|
|
1. **Standard Enhancement**: `Enhance this prompt: {text}` |
|
|
2. **Clean Enhancement**: `Enhance this prompt (no lora): {text}` |
|
|
3. **Technical Enhancement**: `Enhance this prompt (with lora): {text}` |
|
|
4. **Simplification**: `Simplify this prompt: {text}` |
|
|
|
|
|
## Usage |
|
|
|
|
|
```python |
|
|
from optimum.intel import OVModelForSeq2SeqLM |
|
|
from transformers import T5Tokenizer |
|
|
|
|
|
# Load OpenVINO model |
|
|
model = OVModelForSeq2SeqLM.from_pretrained("Mitchins/t5-base-artgen-multi-instruct-OpenVINO") |
|
|
tokenizer = T5Tokenizer.from_pretrained("Mitchins/t5-base-artgen-multi-instruct-OpenVINO") |
|
|
|
|
|
# Example usage |
|
|
text = "woman in red dress" |
|
|
prompt = f"Enhance this prompt (no lora): {text}" |
|
|
|
|
|
inputs = tokenizer(prompt, return_tensors="pt") |
|
|
outputs = model.generate(**inputs, max_length=80) |
|
|
result = tokenizer.decode(outputs[0], skip_special_tokens=True) |
|
|
``` |
|
|
|
|
|
## Performance |
|
|
|
|
|
Optimized for Intel hardware (CPU, integrated GPU, VPU) with significant speedup compared to standard PyTorch inference. |
|
|
|
|
|
## Deployment |
|
|
|
|
|
Perfect for Intel NUC and other Intel-based edge devices. |
|
|
|
