[ComfyUI] [comfy-env] [VRAM] Before offload 'MoGeWrapper_1' (631 MB) -> cpu: 1271 / 24575 MB Capturing execution frames: sharp_basic.json [ComfyUI] [ComfyUI-MoGe2] Health check: ping (timeout=5.0s)... [ComfyUI] [ComfyUI-MoGe2] Health check: failed [ComfyUI] [ComfyUI-MoGe2] Socket unhealthy, restarting worker... [ComfyUI] [comfy-env] Invalidated 1 stale model patchers (will be cleaned up during next unload) [ComfyUI] [comfy-env] moge2-nodes: Library/bin has 236 DLLs [ComfyUI] [ComfyUI-MoGe2] python: C:\Users\ContainerUser\AppData\Local\Programs\comfy-env\.pixi\envs\moge2-nodes\python.exe [ComfyUI] [ComfyUI-MoGe2] sys_path sent to worker: ['C:\\workspaces\\Sharp-0012\\ComfyUI', 'C:\\workspaces\\Sharp-0012\\ComfyUI\\custom_nodes\\ComfyUI-MoGe2', 'C:\\Users\\ContainerUser\\AppData\\Local\\Programs\\comfy-env\\.pixi\\envs\\moge2-nodes\\Lib\\site-packages', 'C:\\workspaces\\Sharp-0012\\ComfyUI\\custom_nodes\\ComfyUI-MoGe2'] [ComfyUI] [SubprocessWorker] is_pixi=True, python=C:\Users\ContainerUser\AppData\Local\Programs\comfy-env\.pixi\envs\moge2-nodes\python.exe [ComfyUI] [SubprocessWorker] launching cmd=['C:\\Users\\ContainerUser\\.pixi\\bin\\pixi.exe', 'run', '--as-is', '--manifest-path', 'C:\\Users\\ContainerUser\\AppData\\Local\\Programs\\comfy-env\\pixi.toml', '-e', 'moge2-nodes', 'python', 'C:\\Users\\ContainerUser\\AppData\\Local\\Temp\\comfyui_pvenv_wfep91u2\\persistent_worker.py', 'tcp://127.0.0.1:55567']... [ComfyUI] [worker:ComfyUI-MoGe2] ERROR: Failed to import comfy_kitchen, Error: No module named 'comfy_kitchen', fp8 and fp4 support will not be available. [ComfyUI] [worker:ComfyUI-MoGe2] WARNING: comfy_kitchen does not support stochastic FP8 rounding, please update comfy_kitchen. [ComfyUI] [SubprocessWorker] call_id=3 sending model_to_device [ComfyUI] [WARNING] Model 'MoGeWrapper_1' gone from worker (restarted?), treating as offloaded [ComfyUI] [comfy-env] [VRAM] After offload 'MoGeWrapper_1' (631 MB): 1271 / 24575 MB [ComfyUI] [comfy-env] [VRAM] Before offload 'RGBGaussianPredictor_1' (2679 MB) -> cpu: 1271 / 24575 MB [ComfyUI] [ComfyUI-Sharp] Health check: ping (timeout=5.0s)... [ComfyUI] [ComfyUI-Sharp] Health check: ok [ComfyUI] [SubprocessWorker] call_id=5 sending model_to_device [ComfyUI] [worker:ComfyUI-Sharp] [VRAM] -290MB (now 2399MB) reserved=7104MB peak=8917MB [ComfyUI] [comfy-env] [VRAM] After offload 'RGBGaussianPredictor_1' (2679 MB): 1271 / 24575 MB [ComfyUI] [INFO] Using RAM pressure cache. Frame 1 saved (frame_000.jpg, t=0.0s) Validating workflow... [ComfyUI] [INFO] === /validate endpoint called === [ComfyUI] [INFO] Received JSON with keys: ['prompt'] [ComfyUI] [INFO] Prompt has 4 nodes Queuing workflow for execution... [ComfyUI] [INFO] Node 1: LoadImage [ComfyUI] [INFO] Inputs: {"image": "beach.jpg"} [ComfyUI] [INFO] Node 2: LoadSharpModel [ComfyUI] [INFO] Inputs: {"precision": "auto"} [ComfyUI] [INFO] Node 5: SharpPredict [ComfyUI] [INFO] Inputs: {"focal_length_mm": 30, "output_prefix": "sharp", "model": ["2", 0], "image": ["1", 0]} [ComfyUI] [INFO] Node 6: PreviewGaussianSharp [ComfyUI] [INFO] Inputs: {"preview_gaussian": "", "ply_path": ["5", 0], "extrinsics": ["5", 1], "intrinsics": ["5", 2]} [ComfyUI] [INFO] Generated prompt_id: c128fe63-3156-4ee4-8b98-9922172251ef [ComfyUI] [INFO] Calling execution.validate_prompt()... [ComfyUI] [INFO] Validation result: valid=True [ComfyUI] [INFO] === Validation PASSED === [ComfyUI] [INFO] got prompt [ComfyUI] [comfy-env] Running LoadSharpModel: precision: 'auto' [ComfyUI] [VRAM] Before LoadSharpModel: 1271 / 24575 MB [ComfyUI] [SubprocessWorker] call_method: nodes.load_model.LoadSharpModel.EXECUTE_NORMALIZED [ComfyUI] [SubprocessWorker] acquired lock, ensuring started... [ComfyUI] [ComfyUI-Sharp] Health check: ping (timeout=5.0s)... [ComfyUI] [ComfyUI-Sharp] Health check: failed [ComfyUI] [ComfyUI-Sharp] Socket unhealthy, restarting worker... [ComfyUI] [comfy-env] Invalidated 1 stale model patchers (will be cleaned up during next unload) [ComfyUI] [comfy-env] sharp-nodes: Library/bin has 565 DLLs [ComfyUI] [ComfyUI-Sharp] python: C:\Users\ContainerUser\AppData\Local\Programs\comfy-env\.pixi\envs\sharp-nodes\python.exe [ComfyUI] [ComfyUI-Sharp] sys_path sent to worker: ['C:\\workspaces\\Sharp-0012\\ComfyUI', 'C:\\workspaces\\Sharp-0012\\ComfyUI\\custom_nodes\\ComfyUI-Sharp', 'C:\\Users\\ContainerUser\\AppData\\Local\\Programs\\comfy-env\\.pixi\\envs\\sharp-nodes\\Lib\\site-packages', 'C:\\workspaces\\Sharp-0012\\ComfyUI\\custom_nodes\\ComfyUI-Sharp'] [ComfyUI] [SubprocessWorker] is_pixi=True, python=C:\Users\ContainerUser\AppData\Local\Programs\comfy-env\.pixi\envs\sharp-nodes\python.exe [ComfyUI] [SubprocessWorker] launching cmd=['C:\\Users\\ContainerUser\\.pixi\\bin\\pixi.exe', 'run', '--as-is', '--manifest-path', 'C:\\Users\\ContainerUser\\AppData\\Local\\Programs\\comfy-env\\pixi.toml', '-e', 'sharp-nodes', 'python', 'C:\\Users\\ContainerUser\\AppData\\Local\\Temp\\comfyui_pvenv_jz5clg4r\\persistent_worker.py', 'tcp://127.0.0.1:52895']... Frame 2 saved (frame_001.jpg, t=2.3s) [capture-periodic] freeze=3ms shot=131ms unfreeze=8ms saved=True [ComfyUI] [worker:ComfyUI-Sharp] ERROR: Failed to import comfy_kitchen, Error: No module named 'comfy_kitchen', fp8 and fp4 support will not be available. [ComfyUI] [worker:ComfyUI-Sharp] WARNING: comfy_kitchen does not support stochastic FP8 rounding, please update comfy_kitchen. [ComfyUI] [SubprocessWorker] worker started/confirmed [ComfyUI] [SubprocessWorker] serializing kwargs to shm... [ComfyUI] [SubprocessWorker] created 0 shm blocks [ComfyUI] [SubprocessWorker] sending request via socket... [ComfyUI] [SubprocessWorker] call_id=6 sending call_method [ComfyUI] [worker:ComfyUI-Sharp] [LoadSharpModel] precision='auto' -> dtype=bf16 (device=cuda:0) [ComfyUI] [worker:ComfyUI-Sharp] [LoadSharpModel] checkpoint already present, reusing (no download): C:\workspaces\Sharp-0012\ComfyUI\models\sharp\sharp_2572gikvuh.pt (2679.6 MB) [ComfyUI] [worker:ComfyUI-Sharp] [LoadSharpModel] checkpoint ready: C:\workspaces\Sharp-0012\ComfyUI\models\sharp\sharp_2572gikvuh.pt (2679.6 MB) [ComfyUI] [worker:ComfyUI-Sharp] [LoadSharpModel] config: precision=auto -> dtype=bf16 [ComfyUI] [SubprocessWorker] got response: ok [ComfyUI] [comfy-env] [VRAM] Before offload 'RGBGaussianPredictor_1' (2679 MB) -> cpu: 1271 / 24575 MB [ComfyUI] [ComfyUI-Sharp] Health check: ping (timeout=5.0s)... [ComfyUI] [ComfyUI-Sharp] Health check: ok [ComfyUI] [SubprocessWorker] call_id=7 sending model_to_device [ComfyUI] [WARNING] Model 'RGBGaussianPredictor_1' gone from worker (restarted?), treating as offloaded [ComfyUI] [comfy-env] [VRAM] After offload 'RGBGaussianPredictor_1' (2679 MB): 1271 / 24575 MB [ComfyUI] [comfy-env] [VRAM] Before offload 'MoGeWrapper_1' (631 MB) -> cpu: 1271 / 24575 MB [ComfyUI] [ComfyUI-MoGe2] Health check: ping (timeout=5.0s)... [ComfyUI] [ComfyUI-MoGe2] Health check: ok [ComfyUI] [SubprocessWorker] call_id=4 sending model_to_device [ComfyUI] [WARNING] Model 'MoGeWrapper_1' gone from worker (restarted?), treating as offloaded [ComfyUI] [comfy-env] [VRAM] After offload 'MoGeWrapper_1' (631 MB): 1271 / 24575 MB [ComfyUI] [comfy-env] LoadSharpModel done (3.81s): result: NodeOutput [ComfyUI] [VRAM] After LoadSharpModel: 1271 / 24575 MB [ComfyUI] [comfy-env] Running SharpPredict: focal_length_mm: 30.0, output_prefix: 'sharp', model: dict, image: torch.float32 [1,1024,1536,3] cpu [ComfyUI] [VRAM] Before SharpPredict: 1271 / 24575 MB [ComfyUI] [SubprocessWorker] call_method: nodes.predict.SharpPredict.EXECUTE_NORMALIZED [ComfyUI] [SubprocessWorker] acquired lock, ensuring started... [ComfyUI] [ComfyUI-Sharp] Health check: ping (timeout=5.0s)... [ComfyUI] [ComfyUI-Sharp] Health check: ok [ComfyUI] [SubprocessWorker] worker started/confirmed [ComfyUI] [comfy-env] PRE-SERIALIZE 'image' shape: torch.Size([1, 1024, 1536, 3]) [ComfyUI] [SubprocessWorker] serializing kwargs to shm... [ComfyUI] [SubprocessWorker] created 0 shm blocks [ComfyUI] [SubprocessWorker] sending request via socket... [ComfyUI] [SubprocessWorker] call_id=8 sending call_method [ComfyUI] [worker:ComfyUI-Sharp] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'manager_path', 'storage_key', 'storage_size'] [ComfyUI] [worker:ComfyUI-Sharp] [comfy-env] DESERIALIZE TensorRef: tensor_size=[1, 1024, 1536, 3] [ComfyUI] [worker:ComfyUI-Sharp] [comfy-env] DESERIALIZED tensor shape: torch.Size([1, 1024, 1536, 3]) Frame 3 saved (frame_002.jpg, t=4.5s) [capture-periodic] freeze=3ms shot=137ms unfreeze=5ms saved=True [capture-loop] iter=50 t=6.1s state={'complete': False, 'error': None, 'executedCount': 0, 'wsState': 1} eval_ms=4 ram=16.8/31.9GB (53%) swap=0.0/5.0GB (0%) cpu=19% browser_rss=0.88GB [ComfyUI] [worker:ComfyUI-Sharp] [LoadSharpModel] loading checkpoint from C:\workspaces\Sharp-0012\ComfyUI\models\sharp\sharp_2572gikvuh.pt (dtype=bf16) [ComfyUI] [worker:ComfyUI-Sharp] [LoadSharpModel] initializing model on meta device... [ComfyUI] [worker:ComfyUI-Sharp] [LoadSharpModel] model ready (torch.bfloat16) [ComfyUI] [SubprocessWorker] callback 'request_vram_budget' from call_id=8 [ComfyUI] [comfy-env] VRAM request: 2.81GB, free before eviction: 24.44GB [ComfyUI] [comfy-env] VRAM after eviction: 24.44GB [ComfyUI] [worker:ComfyUI-Sharp] [VRAM] +314MB (now 314MB) reserved=320MB peak=314MB [ComfyUI] [worker:ComfyUI-Sharp] [VRAM] +2181MB (now 2495MB) reserved=2496MB peak=2495MB [capture-loop] iter=100 t=12.2s state={'complete': False, 'error': None, 'executedCount': 0, 'wsState': 1} eval_ms=8 ram=16.9/31.9GB (53%) swap=0.0/5.0GB (0%) cpu=9% browser_rss=0.87GB [ComfyUI] [worker:ComfyUI-Sharp] [VRAM] +809MB (now 3305MB) reserved=3456MB peak=3314MB [ComfyUI] [worker:ComfyUI-Sharp] [VRAM] +858MB (now 4164MB) reserved=7712MB peak=7053MB [ComfyUI] [worker:ComfyUI-Sharp] [VRAM] +234MB (now 4398MB) reserved=6976MB peak=7053MB [ComfyUI] [SubprocessWorker] callback 'report_progress' from call_id=8 [ComfyUI] [SubprocessWorker] got response: ok [ComfyUI] [comfy-env] Created 1 model patchers: ['RGBGaussianPredictor_1'] [ComfyUI] [comfy-env] SharpPredict done (11.39s): result: NodeOutput [ComfyUI] [VRAM] After SharpPredict: 1271 / 24575 MB [ComfyUI] [VRAM] Loaded: SubprocessModel (2679 MB) [ComfyUI] [comfy-env] Running PreviewGaussianSharp: ply_path: 'C:\\workspaces\\Sharp-0012\\ComfyUI\\output\\sharp_17808..., extrinsics: list[4], intrinsics: list[3] [ComfyUI] [VRAM] Before PreviewGaussianSharp: 1271 / 24575 MB [ComfyUI] [VRAM] Loaded: SubprocessModel (2679 MB) [ComfyUI] [SubprocessWorker] call_method: nodes.preview_gaussian_sharp.PreviewGaussianSharp.EXECUTE_NORMALIZED [ComfyUI] [SubprocessWorker] acquired lock, ensuring started... [ComfyUI] [ComfyUI-GaussianPack] Health check: ping (timeout=5.0s)... [ComfyUI] [ComfyUI-GaussianPack] Health check: ok [ComfyUI] [SubprocessWorker] worker started/confirmed [ComfyUI] [SubprocessWorker] serializing kwargs to shm... [ComfyUI] [SubprocessWorker] created 0 shm blocks [ComfyUI] [SubprocessWorker] sending request via socket... [ComfyUI] [SubprocessWorker] call_id=3 sending call_method [ComfyUI] [SubprocessWorker] got response: ok [ComfyUI] [comfy-env] PreviewGaussianSharp done (0.03s): result: NodeOutput [ComfyUI] [VRAM] After PreviewGaussianSharp: 1271 / 24575 MB [ComfyUI] [VRAM] Loaded: SubprocessModel (2679 MB) [ComfyUI] [INFO] Prompt executed in 15.29 seconds [ComfyUI] [comfy-env] [VRAM] Before offload 'RGBGaussianPredictor_1' (2679 MB) -> cpu: 1271 / 24575 MB [ComfyUI] [ComfyUI-Sharp] Health check: ping (timeout=5.0s)... [ComfyUI] [ComfyUI-Sharp] Health check: ok [ComfyUI] [SubprocessWorker] call_id=9 sending model_to_device Frame 4 saved (frame_003.jpg, t=15.4s) [capture-periodic] freeze=6ms shot=147ms unfreeze=18ms saved=True Node executed (1 total), capturing... [ComfyUI] [worker:ComfyUI-Sharp] [VRAM] -3051MB (now 1347MB) reserved=6976MB peak=7053MB [ComfyUI] [comfy-env] [VRAM] After offload 'RGBGaussianPredictor_1' (2679 MB): 1271 / 24575 MB [ComfyUI] [INFO] Using RAM pressure cache. Frame 5 saved (frame_004.jpg, t=15.9s) [capture-node] freeze=1075ms shot=384ms unfreeze=0ms saved=True Execution complete (t=15.9s) [3d-wait] Phase 1: Waiting for 'Loading 3D Model' text to disappear... [3d-wait] Phase 1 done in 0.0s [3d-wait] Phase 2: iframes=1 [3d-wait] iframe viewer_ready=True (matched 1 console logs) [3d-wait] Unfrozen in 0.004s, rendering for 5s... [3d-wait] JS halted via CDP after 13.8s [3d-wait] Re-frozen in 56.3s [3d-wait] Total 3D wait: 56.3s Captured 5 unique frames over 73.76s [debug] iframes: 1 [debug] iframe[0]: src=http://127.0.0.1:41887/extensions/ComfyUI-GaussianPack/viewer_gaussian_sharp.htm, canvas=1536x1024, rafFrozen=True [timing] trigger_3d_previews: 0.0s Saved high-quality screenshot: sharp_basic_executed.png [timing] final_screenshot: 0.3s Captured 6 video frames