clip_text_model

Model type	clip_text_model
Class	CLIPTextModel
Task	feature-extraction
Source	models/clip.py

Description

CLIP text model with causal attention for feature extraction.

Unlike BERT (bidirectional), CLIP’s text encoder uses causal (triangular) attention masking. Outputs last_hidden_state.

Usage

mobius build --model <MODEL_ID> output_dir/

from mobius import build

model = build("<MODEL_ID>")