Skip to content

fix(ci): skip the fused encoder where triton is absent #13

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
jdsika merged 1 commit into from
Nov 11, 2026
Merged

fix(ci): skip the fused encoder where triton is absent #13

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions notebooks/05_inference.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -48,6 +48,7 @@
N = env.num_agents

device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
config["policy"]["fused_slot_encoder"] = config["policy"]["fused_slot_encoder"] and device.type == "cuda"
policy = DrivePolicy(env, **config["policy"]).to(device)

if CHECKPOINT_PATH:
Expand Down
1 change: 1 addition & 0 deletions tests/eval/test_eval.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -877,6 +877,7 @@ def test_policy_builds_from_the_released_driver_env(sdc_replay_map_dir):
with patch.object(sys, "argv", ["pufferl.py"]):
args = pufferl.load_config("puffer_drive")
args["train"]["device"] = "cpu"
args["policy"]["fused_slot_encoder"] = False # the fused encoder needs triton, which is CUDA-only
env_kwargs = {
"num_agents": SDC_SCENARIO_COUNT,
"num_maps": SDC_SCENARIO_COUNT,
Expand Down
3 changes: 2 additions & 1 deletion tests/unit_tests/test_action_selection.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -201,7 +201,8 @@ def test_drive_mean_action_is_exact_in_physical_units(monkeypatch):
action_type="continuous",
)
try:
policy = ocean_torch.Drive(env, **args["policy"])
# the fused encoder is a CUDA/triton kernel and plays no part in the action table
policy = ocean_torch.Drive(env, **{**args["policy"], "fused_slot_encoder": False})

num_lat = len(binding.JERK_LAT)
lat_zero_idx = binding.JERK_LAT.index(0.0)
Expand Down
Loading