google-cloud-vertex
Use Google Cloud Vertex AI and the Gemini API for text generation, embeddings, multimodal inputs, model tuning, and deployment via Python SDK. Use when "vertex ai", "gemini api", "google cloud ai", "embeddings vertex", "imagen", "gemini model", "google ai studio", "vertex deploy".
소스 정보
- 저장소
- Umbral-Bot/umbral-agent-stack
- 최근 소스 활동
- 2026년 3월 23일 23:30
- 감지된 SKILL.md 언어
- 스페인어
- 스타
- 0
- 포크
- 0
설치 방법
기본적으로 소스를 먼저 확인하는 Prompt가 선택됩니다. 직접 명령으로 전환하거나 로컬 사본을 다운로드할 수도 있습니다.
소스 파일 검토
설치 여부를 결정하기 전에 SKILL.md와 SkillsMP에 표시된 보조 파일을 읽어 보세요.
SKILL.md 표시 중
SKILL.md
소스 지침 · 읽기 전용 미리보기- name
- google-cloud-vertex
- description
- Use Google Cloud Vertex AI and the Gemini API for text generation, embeddings, multimodal inputs, model tuning, and deployment via Python SDK. Use when "vertex ai", "gemini api", "google cloud ai", "embeddings vertex", "imagen", "gemini model", "google ai studio", "vertex deploy".
- metadata
- {"openclaw":{"emoji":"💎","requires":{"env":"[Truncated]"}}}
# Google Cloud / Vertex AI Skill
Generar texto, embeddings e imágenes con Gemini y Vertex AI, tanto desde Google AI Studio (API key) como desde Vertex AI (proyecto GCP).
## Requisitos
| Variable | Descripción |
|----------|-------------|
| `GOOGLE_API_KEY_RICK_UMBRAL` | API key de Google AI Studio |
| `GOOGLE_CLOUD_PROJECT_RICK_UMBRAL` | Proyecto GCP para Vertex AI |
### Instalación
```bash
pip install google-genai google-cloud-aiplatform
gcloud auth application-default login
gcloud config set project $GOOGLE_CLOUD_PROJECT_RICK_UMBRAL
```
> **Nota:** `google-cloud-aiplatform` será deprecado en mayo 2026. Google recomienda migrar a `google-genai`.
## 1. Google AI Studio — Gemini con API key
La forma más rápida de usar Gemini. No requiere proyecto GCP.
```python
from google import genai
client = genai.Client(api_key=GOOGLE_API_KEY)
response = client.models.generate_content(
model="gemini-2.5-flash",
contents="Explicá qué es Vertex AI en 3 líneas.",
)
print(response.text)
```
### Streaming
```python
for chunk in client.models.generate_content_stream(
model="gemini-2.5-flash",
contents="Escribí un poema sobre la nube.",
):
print(chunk.text, end="")
```
## 2. Vertex AI — Gemini con proyecto GCP
Para uso enterprise con IAM, VPC, logging y billing por proyecto.
```python
from google import genai
client = genai.Client(
vertexai=True,
project="mi-proyecto",
location="us-central1",
)
response = client.models.generate_content(
model="gemini-2.5-pro",
contents="Analizá este dataset de ventas.",
)
print(response.text)
```
### SDK Legacy (google-cloud-aiplatform)
```python
import vertexai
from vertexai.generative_models import GenerativeModel
vertexai.init(project="mi-proyecto", location="us-central1")
model = GenerativeModel("gemini-2.5-pro")
response = model.generate_content("Tu prompt aquí")
print(response.text)
```
## 3. Embeddings
### Con google-genai
```python
result = client.models.embed_content(
model="gemini-embedding-001",
contents="Texto a embeddear",
)
vector = result.embeddings[0].values # list[float], hasta 3072 dims
```
### Con Vertex AI SDK legacy
```python
from vertexai.language_models import TextEmbeddingModel
model = TextEmbeddingModel.from_pretrained("text-embedding-005")
embeddings = model.get_embeddings(["Texto a embeddear"])
vector = embeddings[0].values # 768 dims
```
### Modelos de embedding disponibles
| Modelo | Dims | Notas |
|--------|------|-------|
| `gemini-embedding-001` | 3072 | State-of-the-art, multilingüe |
| `text-embedding-005` | 768 | Inglés/código |
| `text-multilingual-embedding-002` | 768 | Multilingüe |
## 4. Multimodal — Imagen + Texto
```python
import PIL.Image
img = PIL.Image.open("plano.png")
response = client.models.generate_content(
model="gemini-2.5-pro",
contents=[img, "Describí qué ves en esta imagen."],
)
print(response.text)
```
## 5. Imagen — Generación de imágenes
Task del Worker relacionada: `google.image.generate`
```python
from vertexai.preview.vision_models import ImageGenerationModel
model = ImageGenerationModel.from_pretrained("imagen-3.0-generate-002")
images = model.generate_images(
prompt="Logo minimalista de un agente AI",
number_of_images=1,
aspect_ratio="1:1",
)
images[0].save("logo.png")
```
## 6. gcloud CLI — Comandos útiles
```bash
gcloud ai models list --region=us-central1
gcloud ai endpoints list --region=us-central1
gcloud ai endpoints predict ENDPOINT_ID --region=us-central1 --json-request=input.json
gcloud services enable aiplatform.googleapis.com
```
## Modelos Gemini disponibles
| Modelo | Caso de uso |
|--------|-------------|
| `gemini-2.5-pro` | Razonamiento complejo, coding |
| `gemini-2.5-flash` | Rápido y barato, tareas simples |
| `gemini-2.5-flash-lite` | Ultra-rápido, latencia mínima |
## Notas
- Rick ya tiene `GOOGLE_API_KEY_RICK_UMBRAL` y `GOOGLE_CLOUD_PROJECT_RICK_UMBRAL` configurados.
- AI Studio es para prototyping rápido; Vertex AI es para producción con controles enterprise.
- El alias `gemini_vertex` en `llm.generate` de Rick enruta automáticamente a Vertex AI.
- Docs: https://cloud.google.com/vertex-ai/docs | https://ai.google.dev/gemini-api/docs
GitHub에서 보기