stabilityai
/

stable-audio-3-optimized

audio-generation

Model card Files Files and versions

stable-audio-3-optimized / tensorRT /sm_90

Commit History

Add FP32 SAME-L + SAME-S TRT decoders

773ac48

Cortexelus Claude Opus 4.7 (1M context) commited on Jun 3

Decoder ONNX: clip+scale in FP32 before int32 cast

a429ffa

Cortexelus Claude Opus 4.7 (1M context) commited on Jun 3

DiT engines: replace BF16 with FP16-mixed (FP32 islands) + add FP32 variants

97651ef

Cortexelus Claude Opus 4.7 (1M context) commited on Jun 2

DiT TRT engines: rebuild with profile min=1 (sm_90)

5f27021

Cortexelus Claude Opus 4.7 (1M context) commited on May 29

T5Gemma: FP16-mixed (FP32 attention island) — fixes BF16 numerical bug

23f5c64

Cortexelus commited on May 28

Decoder TRT engines: PCM-baked-in (sm_90)

106e6fd

Cortexelus commited on May 28

Drop L-range suffix from SAME-L decoder filename

633bd26

Cortexelus commited on May 20

Reorganize tensorRT/ into per-arch subdirs; rebuild DiT engines with conditioner baked in

08c1abe

Cortexelus commited on May 20