From 1b68c2c0a0ab6009e5855b1eb5bdec6097d3d544 Mon Sep 17 00:00:00 2001 From: Daniel Han Date: Fri, 17 Jul 2026 13:08:53 +0000 Subject: [PATCH] Pin the measured HiDream quant verdict in tests int8 and fp8 both engage and render cleanly on this family, including short prompts on int8: the routed MoE expert Linears only ever see the concatenated image+text stream (M >> 16), so the torch._int_mm minimum never binds. Pin the no-deny / no-extra-excludes state so a future edit cannot silently regress the measured behavior. --- .../tests/test_diffusion_more_families.py | 16 ++++++++++++++++ 1 file changed, 16 insertions(+) diff --git a/studio/backend/tests/test_diffusion_more_families.py b/studio/backend/tests/test_diffusion_more_families.py index afd0f3df7c..b9515df65d 100644 --- a/studio/backend/tests/test_diffusion_more_families.py +++ b/studio/backend/tests/test_diffusion_more_families.py @@ -287,6 +287,22 @@ def test_ideogram4_memory_table_counts_both_dits(): assert text_encoders_gb > 5.0 +def test_hidream_quant_schemes_not_denied_and_no_extra_excludes(): + # Measured on a B200 (outputs/hidream_smoke): int8 and fp8 both engage and render + # cleanly, including a 2-3 token prompt on int8 -- the routed MoE expert Linears + # only ever see the concatenated image+text token stream (M >> 16), so the + # torch._int_mm minimum never binds and no family exclude tokens are needed. + from core.inference.diffusion_transformer_quant import ( + _FAMILY_SCHEME_DENY, + _INT8_EXCLUDE_NAME_TOKENS, + exclude_tokens_for_scheme, + ) + + assert "hidream-i1" not in _FAMILY_SCHEME_DENY + assert exclude_tokens_for_scheme("int8", "hidream-i1") == _INT8_EXCLUDE_NAME_TOKENS + assert exclude_tokens_for_scheme("fp8", "hidream-i1") == () + + # ── structured exclusions ──────────────────────────────────────────────────── def test_hunyuanimage_is_excluded_with_reason(): reason = excluded_model_reason("tencent/HunyuanImage-3.0")