Ideogram 40
1 mentions across 0 people
All mentions
Unknown speaker
Recommendedpaper · 2026-06-11
“We quantize Ideogram 4.0 - a 9.3B flow-matching diffusion transformer (DiT), shipped as two separate-weight copies of a single-stream 34-layer backbone for classifier-free guidance and conditioned by a Qwen3-VL-8B encoder - for Ampere RTX 3090 GPUs, which lack FP8 tensor cores.”
INT8 Quantization Matches FP8 Quality for 9.3B Diffusion Transformers on Consume ↗