13 days ago
-
Julia-2: julia2_audio_v1 (text = base_v2 byte-identical, + pruned audio tower, stem p9), int8/int4 variants, lean runtime 0.1.0 (audio, LiveSession, modalities), card with speech, live calls and variants
Julia Klee
-
Card: second generation, first-to-second-generation section, images in the state, trimmed; runtime: image input
Julia Klee
-
Remove compiled Python files; add .gitignore
Julia Klee
-
Julia-2 draft: frozen text checkpoint base_v2, draft model card, runtime, metadata
Julia Klee