ceos/unsloth
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Pin the accelerator probes in the DiT family-metadata tests

Browse source 
Six tests read family_train_infos() (or a start preflight) without pinning the
host probes, so they only held on a machine with a bf16 accelerator: on a
GPU-less runner the DiT gate empties precision_modes, turns supports_compile
off, and replaces any other preflight message with the no-accelerator note, and
all six failed there. A conftest fixture pins both probes for exactly those
tests, so they assert the family metadata they are about on every host. The
gate's own CPU-only behaviour keeps its dedicated tests.
This commit is contained in:
Daniel Han 2026-07-26 21:41:28 +00:00
commit 4e32ef6484
5 changed files with 23 additions and 6 deletions
Download patch file Download diff file Expand all files Collapse all files

2
studio/backend/tests/test_diffusion_training.py
View file

@ -1568,7 +1568,7 @@ def test_info_lists_trainable_families(client):
assert families["z-image"]["defaults"]["resolution"] == 768
def test_start_gated_base_without_access_is_400_and_keeps_gpu(client, monkeypatch):
def test_start_gated_base_without_access_is_400_and_keeps_gpu(client, monkeypatch, dit_train_host):
# A gated FLUX base with no valid token must 400 from the HEAD preflight BEFORE the GPU residents
# are freed, so a doomed start never evicts the user's loaded model.
import urllib.error