Skill UI

Bridges frozen vision encoders and LLMs for high-quality image captioning, VQA, retrieval, and multimodal chat, enabling zero-shot reasoning without fine-tuning and offering efficient Q-Former training.

CLIP enables general-purpose vision-language tasks through zero-shot image classification, similarity scoring, semantic search, and content moderation using pretrained contrastive embeddings without fine-tuning.

Provides Meta AI’s Segment Anything Model for zero-shot image segmentation via interactive prompts or automatic mask generation, enabling rapid dataset creation and integration in vision pipelines.

This skill provides zero-shot time series forecasting using Google's TimesFM foundation model. It forecasts any univariate time series (like sales, sensor data, or weather) without requiring custom model training. The model delivers point forecasts along with calibrated prediction intervals (quantiles), offering a robust and probabilistic view of future trends. It is ideal for rapid, reliable forecasting across diverse time-based datasets.

This skill utilizes the TimesFM foundation model for zero-shot forecasting of any univariate time series. It predicts future values with calibrated probabilistic intervals without requiring custom model training. It includes a mandatory preflight system checker to ensure adequate RAM, GPU, and disk resources for safe, local inference across various data types (sales, sensor, energy).