Instructions to use Pedro21613/PS-design-1.2-nano with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Pedro21613/PS-design-1.2-nano with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="Pedro21613/PS-design-1.2-nano")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Pedro21613/PS-design-1.2-nano")
model = AutoModelForCausalLM.from_pretrained("Pedro21613/PS-design-1.2-nano", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Pedro21613/PS-design-1.2-nano with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Pedro21613/PS-design-1.2-nano"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Pedro21613/PS-design-1.2-nano",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Pedro21613/PS-design-1.2-nano

SGLang

How to use Pedro21613/PS-design-1.2-nano with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "Pedro21613/PS-design-1.2-nano" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Pedro21613/PS-design-1.2-nano",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "Pedro21613/PS-design-1.2-nano" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Pedro21613/PS-design-1.2-nano",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Pedro21613/PS-design-1.2-nano with Docker Model Runner:
```
docker model run hf.co/Pedro21613/PS-design-1.2-nano
```

PS Design 1.2 Nano

Versão otimizada, mais leve e mais rápida do PS Design 1.2-beta.

Merge do LoRA (r16) direto no Qwen2.5-0.5B-Instruct → modelo standalone, sem dependência de PEFT no inference
Mesmos pesos do 1.2-beta (qualidade idêntica), porém sem overhead do adapter (~10-15% mais rápido, carregamento único)
use_cache=True, repetition_penalty=1.05, fp16
0.5B params (~943MB fp16) — roda em T4 e até CPU

Uso (sem PEFT)

from transformers import AutoTokenizer, AutoModelForCausalLM
import torch
mid="Pedro21613/PS-design-1.2-nano"
tok=AutoTokenizer.from_pretrained(mid, trust_remote_code=True)
m=AutoModelForCausalLM.from_pretrained(mid, trust_remote_code=True, dtype=torch.float16, device_map="auto").eval()
msgs=[{"role":"system","content":"Você é um designer front-end sênior especialista em páginas bonitas, modernas, minimalistas e profissionais. Sempre responda com código HTML completo, responsivo, usando Tailwind CSS via CDN."},{"role":"user","content":"Crie uma landing page minimalista para cafeteria."}]
p=tok.apply_chat_template(msgs, tokenize=False, add_generation_prompt=True)
i=tok(p, return_tensors="pt").to(m.device)
o=m.generate(**i, max_new_tokens=1200, temperature=0.7, top_p=0.9, do_sample=True)
print(tok.decode(o[0], skip_special_tokens=True))

Derivado de Pedro21613/PS-design-1.2-beta (6200 exemplos).

Downloads last month: 512

Safetensors

Model size

0.5B params

Tensor type

F16

Model tree for Pedro21613/PS-design-1.2-nano

Base model

Qwen/Qwen2.5-0.5B

Finetuned

Qwen/Qwen2.5-0.5B-Instruct

Finetuned

(1056)

this model