Merge branch 'fix/05-inference-yaml-params' into feature/auto-pipeline

This commit is contained in:
Ubuntu
2026-05-12 22:46:30 +00:00

View File

@@ -32,11 +32,24 @@ import sys
import time import time
from pathlib import Path from pathlib import Path
import yaml
sys.path.insert(0, str(Path(__file__).parent.parent)) sys.path.insert(0, str(Path(__file__).parent.parent))
from orchestrator.db import init_db, get_conn, upsert_job, record_metric, now_iso from orchestrator.db import init_db, get_conn, upsert_job, record_metric, now_iso
PIPELINE_BASE = Path(os.environ.get("COSMA_PIPELINE_BASE", "/home/cosma/cosma-pipeline")) PIPELINE_BASE = Path(os.environ.get("COSMA_PIPELINE_BASE", "/home/cosma/cosma-pipeline"))
def _load_inference_cfg() -> dict:
"""Load inference params from thresholds.yaml, with sane defaults."""
cfg_path = Path(__file__).parent.parent / "config" / "thresholds.yaml"
try:
data = yaml.safe_load(cfg_path.read_text())
return data.get("inference", {})
except Exception:
return {}
_INF_CFG = _load_inference_cfg()
WORKERS = { WORKERS = {
".84": { ".84": {
"host": "192.168.0.84", "host": "192.168.0.84",
@@ -146,19 +159,37 @@ def run_inference(frames_dir: Path, worker_key: str, mission_name: str,
return metrics return metrics
print(f" [05] rsync done") print(f" [05] rsync done")
# Step 2: build demo.py command # Step 2: build demo.py command -- params from thresholds.yaml[inference]
checkpoint = f"{w['ai_dir']}/checkpoints/lingbot-map/lingbot-map.pt" checkpoint = f"{w['ai_dir']}/checkpoints/lingbot-map/lingbot-map.pt"
inf_mode = _INF_CFG.get("mode", "streaming")
conf_thr = _INF_CFG.get("ply_conf_threshold", 1.5)
kf_interval = _INF_CFG.get("keyframe_interval", 1)
max_frames = _INF_CFG.get("max_frame_num", 1024)
if inf_mode == "windowed":
window_size = _INF_CFG.get("window_size", 64)
overlap_size = _INF_CFG.get("overlap_size", 16)
mode_flags = (
f"--mode windowed "
f"--window_size {window_size} "
f"--overlap_size {overlap_size} "
)
else: # streaming (default, validated GX049839_v2 146M pts)
mode_flags = (
f"--mode streaming "
f"--keyframe_interval {kf_interval} "
f"--max_frame_num {max_frames} "
)
demo_cmd = ( demo_cmd = (
f"cd {w['ai_dir']} && " f"cd {w['ai_dir']} && "
f"{w['venv']} demo.py " f"{w['venv']} demo.py "
f"--model_path {checkpoint} " f"--model_path {checkpoint} "
f"--image_folder {worker_frames} " f"--image_folder {worker_frames} "
f"--mode windowed " f"{mode_flags}"
f"--window_size 64 " f"--ply_conf_threshold {conf_thr} "
f"--overlap_size 16 "
f"--save_ply {ply_remote} " f"--save_ply {ply_remote} "
f"--save_poses {npz_remote} " f"--save_poses {npz_remote} "
f"--use_sdpa " f"--use_sdpa "
f"--offload_to_cpu "
f"2>&1" f"2>&1"
) )