[00:00] Crash dump logging enabled: /logs/DepthAnythingV3-0307/dev/linux-gpu/crash_dump.log [00:00] [00:00] [1/5] SYNTAX [00:00] ---------------------------------------- [00:00] [DEBUG] server=None, server_url=None, api=None [00:00] Found pyproject.toml (modern format) [00:00] Found requirements.txt (legacy format) [00:00] Unicode check: OK (all characters cp1252-safe) [00:00] Forbidden patterns check: OK [00:00] [SYNTAX] PASSED [00:00] [00:00] [2/5] INSTALL [00:00] ---------------------------------------- [00:00] [DEBUG] server=None, server_url=None, api=None [00:00] Setting up ComfyUI... [00:00] Creating virtual environment... [00:00] Running: uv venv /workspaces/DepthAnythingV3-0307/.venv --python 3.11 [00:00] Cloning ComfyUI (latest)... [00:00] Running: git clone --depth 1 https://github.com/comfyanonymous/ComfyUI.git /workspaces/DepthAnythingV3-0307/ComfyUI [00:02] Installing ComfyUI requirements... [00:02] Running: uv pip install --python /workspaces/DepthAnythingV3-0307/.venv/bin/python --index-url https://download.pytorch.org/whl/cu128 --extra-index-url https://pypi.org/simple --index-strategy unsafe-best-match -r /workspaces/DepthAnythingV3-0307/ComfyUI/requirements.txt [01:12] Installing custom node... [01:13] Copying ComfyUI-DepthAnythingV3 to custom_nodes/... [01:13] Installing node requirements... [01:13] Running: uv pip install --python /workspaces/DepthAnythingV3-0307/.venv/bin/python --index-url https://download.pytorch.org/whl/cu128 --extra-index-url https://pypi.org/simple --index-strategy unsafe-best-match -r /workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/requirements.txt [01:15] Running install.py... [01:15] Running: /workspaces/DepthAnythingV3-0307/.venv/bin/python /workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/install.py [01:16] [01:16] Installing 3 node dependencies... [01:16] Cloning ComfyUI-VideoHelperSuite... [01:17] Installing requirements for ComfyUI-VideoHelperSuite... [01:19] Cloning ComfyUI-KJNodes... [01:21] Installing requirements for ComfyUI-KJNodes... [01:23] Cloning ComfyUI-GeometryPack... [01:28] Installing requirements for ComfyUI-GeometryPack... [01:29] Running install.py for ComfyUI-GeometryPack... [01:32] [comfy-env] Workspace: /workspaces/DepthAnythingV3-0307/ComfyUI/.ce [01:32] [comfy-env] ComfyUI: /workspaces/DepthAnythingV3-0307/ComfyUI [01:32] [comfy-env] Found 2 node config(s): [01:33] - depthanythingv3-nodes <- custom_nodes/ComfyUI-DepthAnythingV3/nodes/comfy-env.toml (python=host) [01:33] - geometrypack-nodes <- custom_nodes/ComfyUI-GeometryPack/nodes/comfy-env.toml (python=host) [01:33] [comfy-env] Bootstrap interpreter has python 3.11 + torch 2.11.0 (cu128). Looking for cuda-only wheels matching cu128/torch2.11/cp311. [01:33] [comfy-env] GPU: Detection: smi [01:33] Driver: 580.126.20 [01:33] CUDA: 13.0 [01:33] Recommended: CUDA 13.0 [01:33] [01:33] GPU 0: NVIDIA GeForce RTX 5060 (sm_120) [Blackwell] 8151MB [01:33] [comfy-env] Comfyui feature -> torch 2.11.0 from https://download.pytorch.org/whl/cu128 (CUDA 12.8 via bootstrap torch) [01:33] [comfy-env] cuda-wheels: 3 package(s) need a matched (cuda, torch, python) combo: ['flash_attn', 'sageattention', 'cumesh'] [01:33] [comfy-env] cuda-wheels tier 1 (bootstrap): probing cu12.8/torch2.11/cp311 [01:34] [cuda-wheels] Looking up flash_attn: cu128 torch2.11 cp311 manylinux linux [01:34] [cuda-wheels] https://pozzettiandrea.github.io/cuda-wheels/v2/flash_attn/: HTTPError: HTTP Error 404: Not Found [01:34] [cuda-wheels] Found: flash_attn-2.8.3+cu128torch2.11-cp311-cp311-manylinux_2_34_x86_64.manylinux_2_35_x86_64.whl [01:34] [cuda-wheels] Looking up sageattention: cu128 torch2.11 cp311 manylinux linux [01:34] [cuda-wheels] Found: sageattention-2.2.0+cu128torch2.11-cp311-cp311-manylinux_2_34_x86_64.manylinux_2_35_x86_64.whl [01:35] [cuda-wheels] Looking up cumesh: cu128 torch2.11 cp311 manylinux linux [01:35] [cuda-wheels] Found: cumesh-0.0.1+cu128torch2.11-cp311-cp311-manylinux_2_35_x86_64.whl [01:35] [comfy-env] cuda-wheels combo: cu12.8/torch2.11 (bootstrap matches; per-node envs will pin to this) [01:36] [comfy-env] Root conda deps from ComfyUI-DepthAnythingV3: ['av', 'ffmpeg'] [01:36] [comfy-env] Root conda deps from ComfyUI-GeometryPack: ['av', 'ffmpeg'] [01:36] [comfy-env] Comfyui feature pin: torch ==2.11.0, torchvision ==0.26.*, torchaudio ==2.11.* [01:36] [comfy-env] Host glibc 2.35 -> comfyui feature system-requirements [01:36] [comfy-env] Root conda deps: ['av', 'ffmpeg'] [01:36] [comfy-env] comfyui311: python 3.11, torch from https://download.pytorch.org/whl/cu128 [01:36] [comfy-env] cuda-wheels combo: per-node cuda features will pin ['torch', 'torchaudio', 'torchvision'] via pypi-options.dependency-overrides (https://download.pytorch.org/whl/cu128) [01:36] [cuda-wheels] Looking up flash_attn: cu128 torch2.11 cp311 manylinux linux [01:36] [cuda-wheels] https://pozzettiandrea.github.io/cuda-wheels/v2/flash_attn/: HTTPError: HTTP Error 404: Not Found [01:36] [cuda-wheels] Found: flash_attn-2.8.3+cu128torch2.11-cp311-cp311-manylinux_2_34_x86_64.manylinux_2_35_x86_64.whl [01:36] [cuda-wheels] Looking up sageattention: cu128 torch2.11 cp311 manylinux linux [01:36] [cuda-wheels] Found: sageattention-2.2.0+cu128torch2.11-cp311-cp311-manylinux_2_34_x86_64.manylinux_2_35_x86_64.whl [01:36] [comfy-env] depthanythingv3-nodes: cuda-wheels deferred for post-pixi install (flash_attn, sageattention) [01:36] [comfy-env] depthanythingv3-nodes: torch override -> torch==2.11.0, torchvision==0.26.*, torchaudio==2.11.* from https://download.pytorch.org/whl/cu128 [01:36] [cuda-wheels] Looking up cumesh: cu128 torch2.11 cp311 manylinux linux [01:36] [cuda-wheels] Found: cumesh-0.0.1+cu128torch2.11-cp311-cp311-manylinux_2_35_x86_64.whl [01:36] [comfy-env] geometrypack-nodes: cuda-wheels deferred for post-pixi install (cumesh) [01:36] [comfy-env] geometrypack-nodes: torch override -> torch==2.11.0, torchvision==0.26.*, torchaudio==2.11.* from https://download.pytorch.org/whl/cu128 [01:36] Generated /workspaces/DepthAnythingV3-0307/ComfyUI/.ce/pixi.toml [01:36] [comfy-env] Installing 2 environment(s): [01:36] - depthanythingv3-nodes (python=host, 9 pypi deps, 2 cuda wheels) [01:36] - geometrypack-nodes (python=host, 13 pypi deps, 1 cuda wheels) [01:36] [comfy-env] Running `pixi install --all`... [01:49] WARN Skipped running the post-link scripts because `run-post-link-scripts` = `false` [01:49] - bin/.librsvg-pre-unlink.sh [01:49] To enable them, run: [01:50] pixi config set --local run-post-link-scripts insecure [01:50] More info: [01:50] https://pixi.sh/latest/reference/pixi_configuration/#run-post-link-scripts [01:54] WARN Skipped running the post-link scripts because `run-post-link-scripts` = `false` [01:54] - bin/.librsvg-pre-unlink.sh [01:54] To enable them, run: [01:54] pixi config set --local run-post-link-scripts insecure [01:54] More info: [01:54] https://pixi.sh/latest/reference/pixi_configuration/#run-post-link-scripts [01:57] WARN Skipped running the post-link scripts because `run-post-link-scripts` = `false` [01:57] - bin/.librsvg-pre-unlink.sh [01:57] To enable them, run: [01:57] pixi config set --local run-post-link-scripts insecure [01:57] More info: [01:57] https://pixi.sh/latest/reference/pixi_configuration/#run-post-link-scripts [03:05] The following environments have been installed: default, comfyui, depthanythingv3-nodes, geometrypack-nodes. [03:05] [comfy-env] Note: env `default` is on disk but no node declares it in this run. Leaving as-is. Remove via `pixi clean --environment default` if intended. [03:05] [comfy-env] depthanythingv3-nodes: installing cuda-wheels with --no-deps (flash_attn, sageattention) [03:05] Using Python 3.11.15 environment at: .pixi/envs/depthanythingv3-nodes [03:07] Resolved 2 packages in 2.00s [03:07] Downloading flash-attn (242.3MiB) [03:07] Downloading sageattention (53.2MiB) [03:08] Downloaded sageattention [03:15] Downloaded flash-attn [03:15] Prepared 2 packages in 8.00s [03:15] Installed 2 packages in 8ms [03:15] + flash-attn==2.8.3+cu128torch2.11 (from https://github.com/PozzettiAndrea/cuda-wheels/releases/download/flash_attn-latest/flash_attn-2.8.3%2Bcu128torch2.11-cp311-cp311-manylinux_2_34_x86_64.manylinux_2_35_x86_64.whl) [03:15] + sageattention==2.2.0+cu128torch2.11 (from https://github.com/PozzettiAndrea/cuda-wheels/releases/download/sageattention-latest/sageattention-2.2.0%2Bcu128torch2.11-cp311-cp311-manylinux_2_34_x86_64.manylinux_2_35_x86_64.whl) [03:15] [comfy-env] geometrypack-nodes: installing cuda-wheels with --no-deps (cumesh) [03:15] Using Python 3.11.15 environment at: .pixi/envs/geometrypack-nodes [03:16] Resolved 1 package in 843ms [03:16] Downloading cumesh (19.5MiB) [03:16] Downloaded cumesh [03:16] Prepared 1 package in 337ms [03:16] Installed 1 package in 9ms [03:16] + cumesh==0.0.1+cu128torch2.11 (from https://github.com/PozzettiAndrea/cuda-wheels/releases/download/cumesh-latest/cumesh-0.0.1%2Bcu128torch2.11-cp311-cp311-manylinux_2_35_x86_64.whl) [03:16] [comfy-env] Install log: /workspaces/DepthAnythingV3-0307/ComfyUI/.ce/install.log [03:17] [03:17] Installation complete! [03:17] [03:17] [requirements] Re-installing main package requirements... [03:17] Installing requirements for ComfyUI-DepthAnythingV3... [03:17] [comfy-env] Workspace: /workspaces/DepthAnythingV3-0307/ComfyUI/.ce [03:17] [comfy-env] ComfyUI: /workspaces/DepthAnythingV3-0307/ComfyUI [03:17] [comfy-env] Found 2 node config(s): [03:17] - depthanythingv3-nodes <- custom_nodes/ComfyUI-DepthAnythingV3/nodes/comfy-env.toml (python=host) [03:17] - geometrypack-nodes <- custom_nodes/ComfyUI-GeometryPack/nodes/comfy-env.toml (python=host) [03:17] [comfy-env] Bootstrap interpreter has python 3.11 + torch 2.11.0 (cu128). Looking for cuda-only wheels matching cu128/torch2.11/cp311. [03:18] [comfy-env] GPU: Detection: smi [03:18] Driver: 580.126.20 [03:18] CUDA: 13.0 [03:18] Recommended: CUDA 13.0 [03:18] [03:18] GPU 0: NVIDIA GeForce RTX 5060 (sm_120) [Blackwell] 8151MB [03:18] [comfy-env] Comfyui feature -> torch 2.11.0 from https://download.pytorch.org/whl/cu128 (CUDA 12.8 via bootstrap torch) [03:18] [comfy-env] cuda-wheels: 3 package(s) need a matched (cuda, torch, python) combo: ['flash_attn', 'sageattention', 'cumesh'] [03:18] [comfy-env] cuda-wheels tier 1 (bootstrap): probing cu12.8/torch2.11/cp311 [03:18] [cuda-wheels] Looking up flash_attn: cu128 torch2.11 cp311 manylinux linux [03:18] [cuda-wheels] https://pozzettiandrea.github.io/cuda-wheels/v2/flash_attn/: HTTPError: HTTP Error 404: Not Found [03:18] [cuda-wheels] Found: flash_attn-2.8.3+cu128torch2.11-cp311-cp311-manylinux_2_34_x86_64.manylinux_2_35_x86_64.whl [03:18] [cuda-wheels] Looking up sageattention: cu128 torch2.11 cp311 manylinux linux [03:18] [cuda-wheels] Found: sageattention-2.2.0+cu128torch2.11-cp311-cp311-manylinux_2_34_x86_64.manylinux_2_35_x86_64.whl [03:18] [cuda-wheels] Looking up cumesh: cu128 torch2.11 cp311 manylinux linux [03:18] [cuda-wheels] Found: cumesh-0.0.1+cu128torch2.11-cp311-cp311-manylinux_2_35_x86_64.whl [03:18] [comfy-env] cuda-wheels combo: cu12.8/torch2.11 (bootstrap matches; per-node envs will pin to this) [03:18] [comfy-env] Root conda deps from ComfyUI-DepthAnythingV3: ['av', 'ffmpeg'] [03:18] [comfy-env] Root conda deps from ComfyUI-GeometryPack: ['av', 'ffmpeg'] [03:18] [comfy-env] Comfyui feature pin: torch ==2.11.0, torchvision ==0.26.*, torchaudio ==2.11.* [03:18] [comfy-env] Host glibc 2.35 -> comfyui feature system-requirements [03:18] [comfy-env] Root conda deps: ['av', 'ffmpeg'] [03:18] [comfy-env] comfyui311: python 3.11, torch from https://download.pytorch.org/whl/cu128 [03:18] [comfy-env] cuda-wheels combo: per-node cuda features will pin ['torch', 'torchaudio', 'torchvision'] via pypi-options.dependency-overrides (https://download.pytorch.org/whl/cu128) [03:18] [cuda-wheels] Looking up flash_attn: cu128 torch2.11 cp311 manylinux linux [03:18] [cuda-wheels] https://pozzettiandrea.github.io/cuda-wheels/v2/flash_attn/: HTTPError: HTTP Error 404: Not Found [03:18] [cuda-wheels] Found: flash_attn-2.8.3+cu128torch2.11-cp311-cp311-manylinux_2_34_x86_64.manylinux_2_35_x86_64.whl [03:18] [cuda-wheels] Looking up sageattention: cu128 torch2.11 cp311 manylinux linux [03:19] [cuda-wheels] Found: sageattention-2.2.0+cu128torch2.11-cp311-cp311-manylinux_2_34_x86_64.manylinux_2_35_x86_64.whl [03:19] [comfy-env] depthanythingv3-nodes: cuda-wheels deferred for post-pixi install (flash_attn, sageattention) [03:19] [comfy-env] depthanythingv3-nodes: torch override -> torch==2.11.0, torchvision==0.26.*, torchaudio==2.11.* from https://download.pytorch.org/whl/cu128 [03:19] [cuda-wheels] Looking up cumesh: cu128 torch2.11 cp311 manylinux linux [03:19] [cuda-wheels] Found: cumesh-0.0.1+cu128torch2.11-cp311-cp311-manylinux_2_35_x86_64.whl [03:19] [comfy-env] geometrypack-nodes: cuda-wheels deferred for post-pixi install (cumesh) [03:19] [comfy-env] geometrypack-nodes: torch override -> torch==2.11.0, torchvision==0.26.*, torchaudio==2.11.* from https://download.pytorch.org/whl/cu128 [03:19] Generated /workspaces/DepthAnythingV3-0307/ComfyUI/.ce/pixi.toml [03:19] [comfy-env] Installing 2 environment(s): [03:19] - depthanythingv3-nodes (python=host, 9 pypi deps, 2 cuda wheels) [03:19] - geometrypack-nodes (python=host, 13 pypi deps, 1 cuda wheels) [03:19] [comfy-env] Running `pixi install --all`... [03:19] The following environments have been installed: default, comfyui, depthanythingv3-nodes, geometrypack-nodes. [03:19] [comfy-env] Note: env `default` is on disk but no node declares it in this run. Leaving as-is. Remove via `pixi clean --environment default` if intended. [03:19] [comfy-env] depthanythingv3-nodes: installing cuda-wheels with --no-deps (flash_attn, sageattention) [03:19] Using Python 3.11.15 environment at: .pixi/envs/depthanythingv3-nodes [03:19] Checked 2 packages in 43ms [03:19] [comfy-env] geometrypack-nodes: installing cuda-wheels with --no-deps (cumesh) [03:19] Using Python 3.11.15 environment at: .pixi/envs/geometrypack-nodes [03:19] Checked 1 package in 100ms [03:19] [comfy-env] Install log: /workspaces/DepthAnythingV3-0307/ComfyUI/.ce/install.log [03:19] [03:19] Installation complete! [03:19] Installing 3 node dependency(ies)... [03:19] ComfyUI-VideoHelperSuite from Kosinkadink/ComfyUI-VideoHelperSuite [03:19] ComfyUI-VideoHelperSuite already exists, skipping... [03:19] ComfyUI-KJNodes from kijai/ComfyUI-KJNodes [03:19] ComfyUI-KJNodes already exists, skipping... [03:19] ComfyUI-GeometryPack from PozzettiAndrea/ComfyUI-GeometryPack [03:19] ComfyUI-GeometryPack already exists, skipping... [03:19] Installing validation endpoint... [03:19] Cloning PozzettiAndrea/ComfyUI-validate-endpoint... [03:19] Running: git clone --depth 1 https://github.com/PozzettiAndrea/ComfyUI-validate-endpoint.git /workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-validate-endpoint [03:20] CUDA packages installed (no mock): flash_attn, sageattention [03:20] [INSTALL] PASSED [03:20] [03:20] [3/5] REGISTRATION [03:20] ---------------------------------------- [03:20] [DEBUG] server=None, server_url=None, api=None [03:20] Starting ComfyUI server... [03:20] Starting ComfyUI server on port 41886... [03:20] Starting ComfyUI server on port 41886... [03:20] Waiting for server to be ready (timeout: 600s)... [03:21] [ComfyUI] setup plugin alembic.autogenerate.schemas [03:21] [ComfyUI] setup plugin alembic.autogenerate.tables [03:21] [ComfyUI] setup plugin alembic.autogenerate.types [03:21] [ComfyUI] setup plugin alembic.autogenerate.constraints [03:21] [ComfyUI] setup plugin alembic.autogenerate.defaults [03:21] [ComfyUI] setup plugin alembic.autogenerate.comments [03:21] [ComfyUI] [comfy-env] ComfyUI-DepthAnythingV3: 1 isolation env(s): [03:21] [ComfyUI] [comfy-env] nodes -> /workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes [03:21] [ComfyUI] [comfy-env] prestartup complete [03:21] [ComfyUI] [comfy-env] ComfyUI-GeometryPack: 1 isolation env(s): [03:21] [ComfyUI] [comfy-env] nodes -> /workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-GeometryPack/nodes [03:21] [ComfyUI] [comfy-env] prestartup complete [03:21] [ComfyUI] [03:21] [ComfyUI] Prestartup times for custom nodes: [03:21] [ComfyUI] 0.1 seconds: /workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-GeometryPack [03:21] [ComfyUI] 0.2 seconds: /workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3 [03:21] [ComfyUI] [03:27] [ComfyUI] WARNING: You need pytorch with cu130 or higher to use optimized CUDA operations. [03:27] [ComfyUI] Found comfy_kitchen backend triton: {'available': True, 'disabled': True, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_mxfp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8']} [03:27] [ComfyUI] Found comfy_kitchen backend cuda: {'available': True, 'disabled': True, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_mxfp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8']} [03:27] [ComfyUI] Found comfy_kitchen backend eager: {'available': True, 'disabled': False, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_mxfp8', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_mxfp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8', 'scaled_mm_mxfp8', 'scaled_mm_nvfp4']} [03:27] [ComfyUI] Checkpoint files will always be loaded safely. [03:27] [ComfyUI] Total VRAM 7704 MB, total RAM 32011 MB [03:27] [ComfyUI] pytorch version: 2.11.0+cu128 [03:27] [ComfyUI] Set vram state to: NORMAL_VRAM [03:27] [ComfyUI] Device: cuda:0 NVIDIA GeForce RTX 5060 : cudaMallocAsync [03:27] [ComfyUI] Using async weight offloading with 2 streams [03:27] [ComfyUI] Enabled pinned memory 28810.0 [03:27] [ComfyUI] Using pytorch attention [03:32] [ComfyUI] aimdo: /project/src-posix/cuda-funchooks.c:52:DEBUG:aimdo_setup_hooks: hooks successfully installed [03:32] [ComfyUI] aimdo: /project/src/control.c:151:INFO:comfy-aimdo inited for GPU: NVIDIA GeForce RTX 5060 (VRAM: 7703 MB) [03:32] [ComfyUI] DynamicVRAM support detected and enabled [03:32] [ComfyUI] Python version: 3.11.15 (main, May 8 2026, 18:33:10) [Clang 22.1.3 ] [03:32] [ComfyUI] ComfyUI version: 0.20.1 [03:32] [ComfyUI] comfy-aimdo version: 0.3.0 [03:32] [ComfyUI] comfy-kitchen version: 0.2.8 [03:32] [ComfyUI] ****** User settings have been changed to be stored on the server instead of browser storage. ****** [03:32] [ComfyUI] ****** For multi-user setups add the --multi-user CLI argument to enable multiple user profiles. ****** [03:32] [ComfyUI] ComfyUI frontend version: 1.43.18 [03:32] [ComfyUI] [Prompt Server] web root: /workspaces/DepthAnythingV3-0307/.venv/lib/python3.11/site-packages/comfyui_frontend_package/static [03:32] [ComfyUI] Asset seeder disabled [03:34] [ComfyUI] [comfy-env] Version: 0.3.78 [03:34] [ComfyUI] [comfy-env] register_nodes: ComfyUI-DepthAnythingV3 [03:34] [ComfyUI] [comfy-env] ComfyUI source dir: /workspaces/DepthAnythingV3-0307/ComfyUI [03:34] [ComfyUI] [comfy-env] Found 1 isolation env(s) [03:34] [ComfyUI] [comfy-env] Scanning nodes metadata (isolation root)... [03:34] [ComfyUI] [comfy-env] Metadata scan: /root/.pixi/bin/pixi run --as-is --manifest-path /workspaces/DepthAnythingV3-0307/ComfyUI/.ce/pixi.toml -e depthanythingv3-nodes python /tmp/comfy_meta_t51fbxca.py /workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3 nodes [03:34] [ComfyUI] [comfy-env] Scan env PATH for nodes: [03:34] [ComfyUI] [comfy-env] [0] /root/.local/bin [03:34] [ComfyUI] [comfy-env] [1] /usr/local/nvidia/bin [03:34] [ComfyUI] [comfy-env] [2] /usr/local/cuda/bin [03:34] [ComfyUI] [comfy-env] [3] /usr/local/sbin [03:34] [ComfyUI] [comfy-env] [4] /usr/local/bin [03:34] [ComfyUI] [comfy-env] [5] /usr/sbin [03:34] [ComfyUI] [comfy-env] [6] /usr/bin [03:34] [ComfyUI] [comfy-env] [7] /sbin [03:34] [ComfyUI] [comfy-env] [8] /bin [03:39] [ComfyUI] [comfy-env] Metadata scan stderr for nodes: [03:39] [ComfyUI] [comfy-env] [meta-scan] importing nodes from /workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3 [03:39] [ComfyUI] [comfy-env] WARNING:root:WARNING: You need pytorch with cu130 or higher to use optimized CUDA operations. [03:39] [ComfyUI] [comfy-env] [meta-scan] import OK [03:39] [ComfyUI] [comfy-env] Scanned nodes: 15 nodes (5.4s) [03:39] [ComfyUI] [comfy-env] Scanned nodes root: 15 nodes (5.4s) [03:39] [ComfyUI] [comfy-env] Imported nodes root: 15 nodes (isolation) [03:39] [ComfyUI] [comfy-env] Registered 15 total nodes [03:39] [ComfyUI] loading... [03:39] [ComfyUI] calling register_nodes [03:39] [ComfyUI] [comfy-env] Version: 0.3.78 [03:39] [ComfyUI] [comfy-env] register_nodes: ComfyUI-GeometryPack [03:39] [ComfyUI] [comfy-env] ComfyUI source dir: /workspaces/DepthAnythingV3-0307/ComfyUI [03:39] [ComfyUI] [comfy-env] Found 1 isolation env(s) [03:39] [ComfyUI] [comfy-env] Scanning nodes metadata (isolation root)... [03:39] [ComfyUI] [comfy-env] Metadata scan: /root/.pixi/bin/pixi run --as-is --manifest-path /workspaces/DepthAnythingV3-0307/ComfyUI/.ce/pixi.toml -e geometrypack-nodes python /tmp/comfy_meta_fm29z6om.py /workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-GeometryPack nodes [03:39] [ComfyUI] [comfy-env] Scan env PATH for nodes: [03:39] [ComfyUI] [comfy-env] [0] /root/.local/bin [03:39] [ComfyUI] [comfy-env] [1] /usr/local/nvidia/bin [03:39] [ComfyUI] [comfy-env] [2] /usr/local/cuda/bin [03:39] [ComfyUI] [comfy-env] [3] /usr/local/sbin [03:39] [ComfyUI] [comfy-env] [4] /usr/local/bin [03:39] [ComfyUI] [comfy-env] [5] /usr/sbin [03:39] [ComfyUI] [comfy-env] [6] /usr/bin [03:39] [ComfyUI] [comfy-env] [7] /sbin [03:39] [ComfyUI] [comfy-env] [8] /bin [03:47] [ComfyUI] [comfy-env] Metadata scan stderr for nodes: [03:47] [ComfyUI] [comfy-env] [meta-scan] importing nodes from /workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-GeometryPack [03:47] [ComfyUI] [comfy-env] WARNING:root:WARNING: You need pytorch with cu130 or higher to use optimized CUDA operations. [03:47] [ComfyUI] [comfy-env] [meta-scan] import OK [03:47] [ComfyUI] [comfy-env] Scanned nodes: 125 nodes (7.7s) [03:47] [ComfyUI] [comfy-env] Scanned nodes root: 125 nodes (7.7s) [03:47] [ComfyUI] [comfy-env] Imported nodes root: 125 nodes (isolation) [03:47] [ComfyUI] [comfy-env] Registered 125 total nodes [03:47] [ComfyUI] [03:47] [ComfyUI] Import times for custom nodes: [03:47] [ComfyUI] 0.0 seconds: /workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-validate-endpoint [03:47] [ComfyUI] 0.0 seconds: /workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/websocket_image_save.py [03:47] [ComfyUI] 0.1 seconds: /workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-VideoHelperSuite [03:47] [ComfyUI] 0.1 seconds: /workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-KJNodes [03:47] [ComfyUI] 5.6 seconds: /workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3 [03:47] [ComfyUI] 7.7 seconds: /workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-GeometryPack [03:47] [ComfyUI] [03:47] [ComfyUI] Context impl SQLiteImpl. [03:47] [ComfyUI] Will assume non-transactional DDL. [03:47] [ComfyUI] Context impl SQLiteImpl. [03:47] [ComfyUI] Will assume non-transactional DDL. [03:47] [ComfyUI] Running upgrade -> 0001_assets, Initial assets schema [03:47] [ComfyUI] Revision ID: 0001_assets [03:47] [ComfyUI] Revises: None [03:47] [ComfyUI] Create Date: 2025-12-10 00:00:00 [03:47] [ComfyUI] Running upgrade 0001_assets -> 0002_merge_to_asset_references, Merge AssetInfo and AssetCacheState into unified asset_references table. [03:47] [ComfyUI] Running upgrade 0002_merge_to_asset_references -> 0003_add_metadata_job_id, Add system_metadata and job_id columns to asset_references. [03:47] [ComfyUI] Change preview_id FK from assets.id to asset_references.id. [03:47] [ComfyUI] Database upgraded from None to 0003_add_metadata_job_id [03:47] [ComfyUI] Starting server [03:47] [ComfyUI] [03:47] [ComfyUI] To see the GUI go to: http://127.0.0.1:41886 [03:48] Server responding, waiting for nodes to load... [04:08] Server is ready! [04:08] Checking for import errors in server logs... [04:08] No import errors detected [04:08] Found 1120 registered nodes [04:08] [REGISTRATION] PASSED [04:08] [04:08] [4/5] INSTANTIATION [04:08] ---------------------------------------- [04:08] [DEBUG] server=, server_url=None, api= [04:08] Testing node constructors... [04:08] All 15 node(s) instantiated successfully! [04:08] [INSTANTIATION] PASSED [04:08] [04:08] [5/5] EXECUTION [04:08] ---------------------------------------- [04:08] [DEBUG] server=, server_url=None, api= [04:08] GPU runner - will execute 7 workflow(s) [04:08] Running 7 workflow(s) (all with videos)... [04:08] Installing screenshot dependencies (playwright, pillow)... [04:10] Packages installed, downloading chromium browser... [04:26] Screenshot dependencies installed successfully [04:26] Starting headless browser... [04:27] Installing Vulkan stack (apt)... [04:27] Vulkan install error: [Errno 2] No such file or directory: 'sudo' [04:27] GPU detected (no Vulkan) -- using ANGLE/OpenGL rendering [04:27] Capturing execution frames: advanced.json [04:28] [ComfyUI] [DEPRECATION WARNING] Detected import of deprecated legacy API: /extensions/core/widgetInputs.js. This is likely caused by a custom node extension using outdated APIs. Please update your extensions or contact the extension author for an updated version. [04:33] Frame 1 saved (frame_000.jpg, t=0.0s) [04:33] Validating workflow... [04:33] [ComfyUI] === /validate endpoint called === [04:33] [ComfyUI] Received JSON with keys: ['prompt'] [04:33] [ComfyUI] Prompt has 11 nodes [04:33] [ComfyUI] Node 2: LoadImage [04:33] [ComfyUI] Inputs: {"image": "bridge.jpg"} [04:33] [ComfyUI] Node 3: DownloadAndLoadDepthAnythingV3Model [04:33] [ComfyUI] Inputs: {"model": "da3_large.safetensors", "precision": "auto", "attention": "auto"} [04:33] [ComfyUI] Node 4: PreviewImage [04:33] Queuing workflow for execution... [04:33] [ComfyUI] Inputs: {"images": ["12", 0]} [04:33] [ComfyUI] Node 6: PreviewImage [04:33] [ComfyUI] Inputs: {"images": ["12", 1]} [04:33] [ComfyUI] Node 7: PreviewImage [04:33] [ComfyUI] Inputs: {"images": ["12", 2]} [04:33] [ComfyUI] Node 8: PreviewImage [04:33] [ComfyUI] Inputs: {"images": ["12", 3]} [04:33] [ComfyUI] Node 9: PreviewAny [04:33] [ComfyUI] Inputs: {"source": ["12", 6]} [04:33] [ComfyUI] Node 10: PreviewAny [04:33] [ComfyUI] Inputs: {"source": ["12", 5]} [04:33] [ComfyUI] Node 11: MaskPreview [04:33] [ComfyUI] Inputs: {"mask": ["12", 7]} [04:33] [ComfyUI] Node 12: DepthAnything_V3 [04:33] [ComfyUI] Inputs: {"normalization_mode": "V2-Style", "resize_method": "resize", "invert_depth": false, "keep_model_size": false, "da3_model": ["3", 0], "images": ["2", 0]} [04:33] [ComfyUI] Node 13: PreviewImage [04:33] [ComfyUI] Inputs: {"images": ["12", 4]} [04:33] [ComfyUI] Generated prompt_id: 27946029-7cfa-47bd-9f14-2a0e0be6d4aa [04:33] [ComfyUI] Calling execution.validate_prompt()... [04:33] [ComfyUI] Validation result: valid=True [04:33] [ComfyUI] === Validation PASSED === [04:33] [ComfyUI] got prompt [04:33] [ComfyUI] [comfy-env] Running DownloadAndLoadDepthAnythingV3Model: model: 'da3_large.safetensors', precision: 'auto', attention: 'auto' [04:33] [ComfyUI] [VRAM] Before DownloadAndLoadDepthAnythingV3Model: 269 / 7703 MB [04:33] [ComfyUI] [comfy-env] SubprocessWorker: /workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/bin/python [04:33] [ComfyUI] [comfy-env] env_vars: {'COMFYUI_BASE': '/workspaces/DepthAnythingV3-0307/ComfyUI'} [04:33] [ComfyUI] [SubprocessWorker] call_method: nodes.load_model.DownloadAndLoadDepthAnythingV3Model.EXECUTE_NORMALIZED [04:33] [ComfyUI] [ComfyUI-DepthAnythingV3] python: /workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/bin/python [04:33] [ComfyUI] [SubprocessWorker] acquired lock, ensuring started... [04:33] [ComfyUI] [ComfyUI-DepthAnythingV3] sys_path sent to worker: ['/workspaces/DepthAnythingV3-0307/ComfyUI', '/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3', '/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/lib/python3.11/site-packages', '/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3'] [04:33] [ComfyUI] [SubprocessWorker] is_pixi=True, python=/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/bin/python [04:33] [ComfyUI] [SubprocessWorker] launching cmd=['/root/.pixi/bin/pixi', 'run', '--as-is', '--manifest-path', '/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/pixi.toml', '-e', 'depthanythingv3-nodes', 'python', '/tmp/comfyui_pvenv_4rmut23q/persistent_worker.py', 'abstract://comfy_worker_6856023eca2b']... [04:33] [ComfyUI] [SubprocessWorker] socket_addr=abstract://comfy_worker_6856023eca2b (abstract, no filesystem path) [04:35] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] WARNING: WARNING: You need pytorch with cu130 or higher to use optimized CUDA operations. [04:35] [ComfyUI] [SubprocessWorker] worker started/confirmed [04:35] [ComfyUI] [SubprocessWorker] serializing kwargs to shm... [04:35] [ComfyUI] [SubprocessWorker] created 0 shm blocks [04:35] [ComfyUI] [SubprocessWorker] sending request via socket... [04:35] [ComfyUI] [SubprocessWorker] call_id=1 sending call_method [04:35] Frame 2 saved (frame_001.jpg, t=2.4s) [04:35] [capture-periodic] freeze=4ms shot=141ms unfreeze=5ms saved=True [04:36] [ComfyUI] Warning: You are sending unauthenticated requests to the HF Hub. Please set a HF_TOKEN to enable higher rate limits and faster downloads. [04:36] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] WARNING: Warning: You are sending unauthenticated requests to the HF Hub. Please set a HF_TOKEN to enable higher rate limits and faster downloads. [04:36] [ComfyUI] [04:38] Frame 3 saved (frame_002.jpg, t=4.6s) [04:38] [capture-periodic] freeze=2ms shot=118ms unfreeze=4ms saved=True [04:39] [capture-loop] iter=50 t=5.7s state={'complete': False, 'error': None, 'executedCount': 0, 'wsState': 1} eval_ms=3 ram=5.6/31.3GB (18%) swap=0.0/32.0GB (0%) cpu=9% browser_rss=0.85GB [04:40] [ComfyUI] model.safetensors: 0%| | 0.00/1.64G [00:00 cpu: 3054 / 7703 MB [05:04] Capturing execution frames: advanced_3d.json [05:04] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [05:04] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ok [05:04] [ComfyUI] [SubprocessWorker] call_id=3 sending model_to_device [05:04] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] -280MB (now 359MB) reserved=2560MB peak=1505MB [05:05] [ComfyUI] [comfy-env] [VRAM] After offload 'DepthAnything3Net_1' (783 MB): 558 / 7703 MB [05:10] Frame 1 saved (frame_000.jpg, t=0.0s) [05:10] Validating workflow... [05:10] [ComfyUI] === /validate endpoint called === [05:10] Queuing workflow for execution... [05:10] [ComfyUI] Received JSON with keys: ['prompt'] [05:10] [ComfyUI] Prompt has 6 nodes [05:10] [ComfyUI] Node 2: LoadImage [05:10] [ComfyUI] Inputs: {"image": "kaffi.jpg"} [05:10] [ComfyUI] Node 3: DownloadAndLoadDepthAnythingV3Model [05:10] [ComfyUI] Inputs: {"model": "da3_large.safetensors", "precision": "auto", "attention": "auto"} [05:10] [ComfyUI] Node 21: DepthAnything_V3 [05:10] [ComfyUI] [ComfyUI-DepthAnythingV3] python: /workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/bin/python [05:10] [ComfyUI] Inputs: {"normalization_mode": "Raw", "resize_method": "resize", "invert_depth": false, "keep_model_size": false, "da3_model": ["3", 0], "images": ["2", 0]} [05:10] [ComfyUI] [ComfyUI-DepthAnythingV3] sys_path sent to worker: ['/workspaces/DepthAnythingV3-0307/ComfyUI', '/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3', '/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/lib/python3.11/site-packages', '/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3'] [05:10] [ComfyUI] Node 22: DA3_ToPointCloud [05:10] [ComfyUI] [SubprocessWorker] is_pixi=True, python=/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/bin/python [05:10] [ComfyUI] Inputs: {"confidence_threshold": 0.2, "downsample": 1, "allow_around_1": 1, "filter_outliers": false, "outlier_percentage": 5, "depth_raw": ["21", 0], "confidence": ["21", 1], "intrinsics": ["21", 6], "sky_ma [05:10] [ComfyUI] [SubprocessWorker] launching cmd=['/root/.pixi/bin/pixi', 'run', '--as-is', '--manifest-path', '/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/pixi.toml', '-e', 'depthanythingv3-nodes', 'python', '/tmp/comfyui_pvenv_4rmut23q/persistent_worker.py', 'abstract://comfy_worker_cbac5d477bb0']... [05:10] [ComfyUI] Node 23: DA3_SavePointCloud [05:10] [ComfyUI] [SubprocessWorker] socket_addr=abstract://comfy_worker_cbac5d477bb0 (abstract, no filesystem path) [05:10] [ComfyUI] Inputs: {"filename_prefix": "pointcloud", "pointcloud": ["22", 0]} [05:10] [ComfyUI] Node 25: DA3_PreviewPointCloud [05:10] [ComfyUI] Inputs: {"file_path": ["23", 0], "color_mode": "RGB", "preview_vtk": ""} [05:10] [ComfyUI] Generated prompt_id: a9ed4e9b-bb81-421f-a566-88d6f0cc5b7b [05:10] [ComfyUI] Calling execution.validate_prompt()... [05:10] [ComfyUI] Validation result: valid=True [05:10] [ComfyUI] === Validation PASSED === [05:10] [ComfyUI] got prompt [05:10] [ComfyUI] [comfy-env] Running DownloadAndLoadDepthAnythingV3Model: model: 'da3_large.safetensors', precision: 'auto', attention: 'auto' [05:10] [ComfyUI] [VRAM] Before DownloadAndLoadDepthAnythingV3Model: 558 / 7703 MB [05:10] [ComfyUI] [SubprocessWorker] call_method: nodes.load_model.DownloadAndLoadDepthAnythingV3Model.EXECUTE_NORMALIZED [05:10] [ComfyUI] [SubprocessWorker] acquired lock, ensuring started... [05:10] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [05:10] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: failed [05:10] [ComfyUI] [ComfyUI-DepthAnythingV3] Socket unhealthy, restarting worker... [05:10] [ComfyUI] [comfy-env] Invalidated 1 stale model patchers (will be cleaned up during next unload) [05:12] Frame 2 saved (frame_001.jpg, t=2.4s) [05:12] [capture-periodic] freeze=1ms shot=139ms unfreeze=4ms saved=True [05:12] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] WARNING: WARNING: You need pytorch with cu130 or higher to use optimized CUDA operations. [05:13] [ComfyUI] [SubprocessWorker] worker started/confirmed [05:13] [ComfyUI] [SubprocessWorker] serializing kwargs to shm... [05:13] [ComfyUI] [SubprocessWorker] created 0 shm blocks [05:13] [ComfyUI] [SubprocessWorker] sending request via socket... [05:13] [ComfyUI] [SubprocessWorker] call_id=4 sending call_method [05:14] [ComfyUI] [SubprocessWorker] got response: ok [05:14] [ComfyUI] [comfy-env] [VRAM] Before offload 'DepthAnything3Net_1' (783 MB) -> cpu: 386 / 7703 MB [05:14] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [05:14] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ok [05:14] [ComfyUI] [SubprocessWorker] call_id=5 sending model_to_device [05:14] [ComfyUI] Model 'DepthAnything3Net_1' gone from worker (restarted?), treating as offloaded [05:14] [ComfyUI] [comfy-env] [VRAM] After offload 'DepthAnything3Net_1' (783 MB): 386 / 7703 MB [05:14] [ComfyUI] [comfy-env] DownloadAndLoadDepthAnythingV3Model done (3.71s): result: NodeOutput [05:14] [ComfyUI] [VRAM] After DownloadAndLoadDepthAnythingV3Model: 386 / 7703 MB [05:14] [ComfyUI] [comfy-env] Running DepthAnything_V3: normalization_mode: 'Raw', resize_method: 'resize', invert_depth: False, keep_model_size: False, da3_model: dict, images: torch.float32 [1,533,800,3] cpu [05:14] [ComfyUI] [VRAM] Before DepthAnything_V3: 386 / 7703 MB [05:14] [ComfyUI] [SubprocessWorker] call_method: nodes.nodes_inference.DepthAnything_V3.EXECUTE_NORMALIZED [05:14] [ComfyUI] [SubprocessWorker] acquired lock, ensuring started... [05:14] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [05:14] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ok [05:14] [ComfyUI] [SubprocessWorker] worker started/confirmed [05:14] [ComfyUI] [comfy-env] PRE-SERIALIZE 'images' shape: torch.Size([1, 533, 800, 3]) [05:14] [ComfyUI] [SubprocessWorker] serializing kwargs to shm... [05:14] [ComfyUI] [SubprocessWorker] created 0 shm blocks [05:14] [ComfyUI] [SubprocessWorker] sending request via socket... [05:14] [ComfyUI] [SubprocessWorker] call_id=6 sending call_method [05:14] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [05:14] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[1, 533, 800, 3] [05:14] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([1, 533, 800, 3]) [05:14] Frame 3 saved (frame_002.jpg, t=4.7s) [05:14] [capture-periodic] freeze=2ms shot=137ms unfreeze=5ms saved=True [05:15] [ComfyUI] [SubprocessWorker] callback 'request_vram_budget' from call_id=6 [05:15] [ComfyUI] [comfy-env] VRAM request: 0.82GB, free before eviction: 7.67GB [05:15] [ComfyUI] [comfy-env] VRAM after eviction: 7.67GB [05:15] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +474MB (now 474MB) reserved=480MB peak=474MB [05:15] [capture-loop] iter=50 t=5.8s state={'complete': False, 'error': None, 'executedCount': 0, 'wsState': 1} eval_ms=2 ram=6.0/31.3GB (19%) swap=0.0/32.0GB (0%) cpu=12% browser_rss=0.92GB [05:16] [ComfyUI] [SubprocessWorker] callback 'report_progress' from call_id=6 [05:16] [ComfyUI] [SubprocessWorker] got response: ok [05:16] [ComfyUI] [comfy-env] Created 1 model patchers: ['DepthAnything3Net_1'] [05:16] [ComfyUI] [comfy-env] DepthAnything_V3 done (2.44s): result: NodeOutput [05:16] [ComfyUI] [VRAM] After DepthAnything_V3: 2380 / 7703 MB [05:16] [ComfyUI] [VRAM] Loaded: SubprocessModel (783 MB) [05:16] [ComfyUI] [comfy-env] Running DA3_ToPointCloud: confidence_threshold: 0.2, downsample: 1, allow_around_1: True, filter_outliers: False, outlier_percentage: 5.0, depth_raw: torch.float32 [1,532,800,3] cpu, confidence: torch.float32 [1,532,800,3] cpu, intrinsics: '{\n "intrinsics": [\n {\n "image_0": [\n ..., sky_mask: torch.float32 [1,532,800] cpu, source_image: torch.float32 [1,532,800,3] cpu [05:16] [ComfyUI] [VRAM] Before DA3_ToPointCloud: 2380 / 7703 MB [05:16] [ComfyUI] [VRAM] Loaded: SubprocessModel (783 MB) [05:16] [ComfyUI] [SubprocessWorker] call_method: nodes.nodes_3d.DA3_ToPointCloud.EXECUTE_NORMALIZED [05:16] [ComfyUI] [SubprocessWorker] acquired lock, ensuring started... [05:16] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [05:16] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ok [05:16] [ComfyUI] [SubprocessWorker] worker started/confirmed [05:16] [ComfyUI] [comfy-env] PRE-SERIALIZE 'depth_raw' shape: torch.Size([1, 532, 800, 3]) [05:16] [ComfyUI] [comfy-env] PRE-SERIALIZE 'confidence' shape: torch.Size([1, 532, 800, 3]) [05:16] [ComfyUI] [comfy-env] PRE-SERIALIZE 'sky_mask' shape: torch.Size([1, 532, 800]) [05:16] [ComfyUI] [comfy-env] PRE-SERIALIZE 'source_image' shape: torch.Size([1, 532, 800, 3]) [05:16] [ComfyUI] [SubprocessWorker] serializing kwargs to shm... [05:16] [ComfyUI] [SubprocessWorker] created 0 shm blocks [05:16] [ComfyUI] [SubprocessWorker] sending request via socket... [05:16] [ComfyUI] [SubprocessWorker] call_id=7 sending call_method [05:16] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +318MB (now 792MB) reserved=1888MB peak=1022MB [05:16] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [05:16] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[1, 532, 800, 3] [05:16] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([1, 532, 800, 3]) [05:16] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [05:16] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[1, 532, 800, 3] [05:16] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([1, 532, 800, 3]) [05:16] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [05:16] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[1, 532, 800] [05:16] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([1, 532, 800]) [05:16] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [05:16] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[1, 532, 800, 3] [05:16] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([1, 532, 800, 3]) [05:16] [ComfyUI] [SubprocessWorker] got response: ok [05:16] [ComfyUI] [comfy-env] DA3_ToPointCloud done (0.48s): result: NodeOutput [05:17] [ComfyUI] [VRAM] After DA3_ToPointCloud: 2380 / 7703 MB [05:17] [ComfyUI] [VRAM] Loaded: SubprocessModel (783 MB) [05:17] [ComfyUI] [comfy-env] Running DA3_SavePointCloud: filename_prefix: 'pointcloud', pointcloud: list[1] [05:17] [ComfyUI] [VRAM] Before DA3_SavePointCloud: 2380 / 7703 MB [05:17] [ComfyUI] [VRAM] Loaded: SubprocessModel (783 MB) [05:17] [ComfyUI] [SubprocessWorker] call_method: nodes.nodes_3d.DA3_SavePointCloud.EXECUTE_NORMALIZED [05:17] [ComfyUI] [SubprocessWorker] acquired lock, ensuring started... [05:17] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [05:17] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ok [05:17] [ComfyUI] [SubprocessWorker] worker started/confirmed [05:17] [ComfyUI] [SubprocessWorker] serializing kwargs to shm... [05:17] [ComfyUI] [SubprocessWorker] created 0 shm blocks [05:17] [ComfyUI] [SubprocessWorker] sending request via socket... [05:17] [ComfyUI] [SubprocessWorker] call_id=8 sending call_method [05:17] Frame 4 saved (frame_003.jpg, t=7.0s) [05:17] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [05:17] [capture-periodic] freeze=2ms shot=127ms unfreeze=4ms saved=True [05:17] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[311395, 3] [05:17] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([311395, 3]) [05:17] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [05:17] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[311395] [05:17] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([311395]) [05:17] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [05:17] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[311395, 3] [05:17] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([311395, 3]) [05:19] [ComfyUI] [SubprocessWorker] got response: ok [05:19] [ComfyUI] [comfy-env] DA3_SavePointCloud done (2.35s): result: NodeOutput [05:19] [ComfyUI] [VRAM] After DA3_SavePointCloud: 2380 / 7703 MB [05:19] [ComfyUI] [VRAM] Loaded: SubprocessModel (783 MB) [05:19] [ComfyUI] [comfy-env] Running DA3_PreviewPointCloud: file_path: '/workspaces/DepthAnythingV3-0307/ComfyUI/output/pointclo..., color_mode: 'RGB' [05:19] [ComfyUI] [VRAM] Before DA3_PreviewPointCloud: 2380 / 7703 MB [05:19] [ComfyUI] [VRAM] Loaded: SubprocessModel (783 MB) [05:19] [ComfyUI] [SubprocessWorker] call_method: nodes.preview_nodes.DA3_PreviewPointCloud.EXECUTE_NORMALIZED [05:19] [ComfyUI] [SubprocessWorker] acquired lock, ensuring started... [05:19] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [05:19] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ok [05:19] [ComfyUI] [SubprocessWorker] worker started/confirmed [05:19] [ComfyUI] [SubprocessWorker] serializing kwargs to shm... [05:19] [ComfyUI] [SubprocessWorker] created 0 shm blocks [05:19] [ComfyUI] [SubprocessWorker] sending request via socket... [05:19] [ComfyUI] [SubprocessWorker] call_id=9 sending call_method [05:19] [ComfyUI] [SubprocessWorker] got response: ok [05:19] [ComfyUI] [comfy-env] DA3_PreviewPointCloud done (0.13s): result: NodeOutput [05:19] [ComfyUI] [VRAM] After DA3_PreviewPointCloud: 2380 / 7703 MB [05:19] [ComfyUI] [VRAM] Loaded: SubprocessModel (783 MB) [05:19] [ComfyUI] Prompt executed in 9.19 seconds [05:19] [ComfyUI] [comfy-env] [VRAM] Before offload 'DepthAnything3Net_1' (783 MB) -> cpu: 2380 / 7703 MB [05:19] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [05:19] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ok [05:19] [ComfyUI] [SubprocessWorker] call_id=10 sending model_to_device [05:19] Node executed (2 total), capturing... [05:19] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] -298MB (now 495MB) reserved=1888MB peak=1022MB [05:19] [ComfyUI] [comfy-env] [VRAM] After offload 'DepthAnything3Net_1' (783 MB): 556 / 7703 MB [05:21] Frame 5 saved (frame_004.jpg, t=9.7s) [05:21] [capture-node] freeze=1213ms shot=85ms unfreeze=0ms saved=True [05:21] Execution complete (t=9.7s) [05:21] Captured 5 unique frames over 11.05s [05:21] [debug] iframes: 1 [05:21] [debug] iframe[0]: src=http://127.0.0.1:41886/extensions/ComfyUI-GeometryPack/viewer_vtk.html?v=1778382, canvas=486x394, rafFrozen=True [05:21] [timing] trigger_3d_previews: 0.0s [05:23] Saved high-quality screenshot: advanced_3d_executed.png [05:23] [timing] final_screenshot: 0.3s [05:23] Capturing execution frames: advanced_3d_multiview.json [05:29] Frame 1 saved (frame_000.jpg, t=0.0s) [05:29] Validating workflow... [05:29] [ComfyUI] === /validate endpoint called === [05:29] Queuing workflow for execution... [05:29] [ComfyUI] Received JSON with keys: ['prompt'] [05:29] [ComfyUI] Prompt has 8 nodes [05:29] [ComfyUI] Node 3: DownloadAndLoadDepthAnythingV3Model [05:29] [ComfyUI] Inputs: {"model": "da3_large.safetensors", "precision": "auto", "attention": "auto"} [05:29] [ComfyUI] Node 21: LoadImageDataSetFromFolder [05:29] [ComfyUI] Inputs: {"folder": "SOH"} [05:29] [ComfyUI] Node 27: PreviewImage [05:29] [ComfyUI] Inputs: {"images": ["21", 0]} [05:29] [ComfyUI] Node 38: DA3_MultiViewPointCloud [05:29] [ComfyUI] Inputs: {"extrinsics": ["39", 4], "intrinsics": ["39", 5], "confidence_threshold": 0.3, "downsample": 4, "use_icp": false, "allow_around_1": false, "filter_outliers": true, "outlier_percentage": 5, "depths": [05:29] [ComfyUI] Node 39: DepthAnythingV3_MultiView [05:29] [ComfyUI] Inputs: {"normalization_mode": "Raw", "resize_method": "resize", "invert_depth": false, "keep_model_size": false, "da3_model": ["3", 0], "images": ["21", 0]} [05:29] [ComfyUI] Node 44: DA3_SavePointCloud [05:29] [ComfyUI] Inputs: {"filename_prefix": "pointcloud", "pointcloud": ["38", 0]} [05:29] [ComfyUI] Node 45: PreviewAny [05:29] [ComfyUI] Inputs: {"source": ["44", 0]} [05:29] [ComfyUI] Node 48: DA3_PreviewPointCloud [05:29] [ComfyUI] Inputs: {"file_path": ["44", 0], "color_mode": "RGB", "preview_vtk": ""} [05:29] [ComfyUI] Generated prompt_id: f236930d-957b-463f-aba2-6df815d1359d [05:29] [ComfyUI] Calling execution.validate_prompt()... [05:29] [ComfyUI] Validation result: valid=True [05:29] [ComfyUI] === Validation PASSED === [05:29] [ComfyUI] got prompt [05:29] [ComfyUI] [comfy-env] Running DownloadAndLoadDepthAnythingV3Model: model: 'da3_large.safetensors', precision: 'auto', attention: 'auto' [05:29] [ComfyUI] [ComfyUI-DepthAnythingV3] python: /workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/bin/python [05:29] [ComfyUI] [VRAM] Before DownloadAndLoadDepthAnythingV3Model: 556 / 7703 MB [05:29] [ComfyUI] [ComfyUI-DepthAnythingV3] sys_path sent to worker: ['/workspaces/DepthAnythingV3-0307/ComfyUI', '/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3', '/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/lib/python3.11/site-packages', '/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3'] [05:29] Node executed (1 total), capturing... [05:29] [ComfyUI] [SubprocessWorker] call_method: nodes.load_model.DownloadAndLoadDepthAnythingV3Model.EXECUTE_NORMALIZED [05:29] [ComfyUI] [SubprocessWorker] is_pixi=True, python=/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/bin/python [05:29] [ComfyUI] [SubprocessWorker] acquired lock, ensuring started... [05:29] [ComfyUI] [SubprocessWorker] launching cmd=['/root/.pixi/bin/pixi', 'run', '--as-is', '--manifest-path', '/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/pixi.toml', '-e', 'depthanythingv3-nodes', 'python', '/tmp/comfyui_pvenv_4rmut23q/persistent_worker.py', 'abstract://comfy_worker_3e771f8b19e3']... [05:29] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [05:29] [ComfyUI] [SubprocessWorker] socket_addr=abstract://comfy_worker_3e771f8b19e3 (abstract, no filesystem path) [05:29] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: failed [05:29] [ComfyUI] [ComfyUI-DepthAnythingV3] Socket unhealthy, restarting worker... [05:29] [ComfyUI] [comfy-env] Invalidated 1 stale model patchers (will be cleaned up during next unload) [05:30] Frame 2 saved (frame_001.jpg, t=0.6s) [05:30] [capture-node] freeze=2ms shot=137ms unfreeze=5ms saved=True [05:31] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] WARNING: WARNING: You need pytorch with cu130 or higher to use optimized CUDA operations. [05:32] [ComfyUI] [SubprocessWorker] worker started/confirmed [05:32] [ComfyUI] [SubprocessWorker] serializing kwargs to shm... [05:32] [ComfyUI] [SubprocessWorker] created 0 shm blocks [05:32] [ComfyUI] [SubprocessWorker] sending request via socket... [05:32] [ComfyUI] [SubprocessWorker] call_id=11 sending call_method [05:32] [ComfyUI] [SubprocessWorker] got response: ok [05:32] [ComfyUI] [comfy-env] [VRAM] Before offload 'DepthAnything3Net_1' (783 MB) -> cpu: 386 / 7703 MB [05:32] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [05:32] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ok [05:32] [ComfyUI] [SubprocessWorker] call_id=12 sending model_to_device [05:32] [ComfyUI] Model 'DepthAnything3Net_1' gone from worker (restarted?), treating as offloaded [05:32] [ComfyUI] [comfy-env] [VRAM] After offload 'DepthAnything3Net_1' (783 MB): 386 / 7703 MB [05:32] [ComfyUI] [comfy-env] DownloadAndLoadDepthAnythingV3Model done (2.68s): result: NodeOutput [05:32] [ComfyUI] [VRAM] After DownloadAndLoadDepthAnythingV3Model: 386 / 7703 MB [05:32] [ComfyUI] [comfy-env] Running DepthAnythingV3_MultiView: normalization_mode: 'Raw', resize_method: 'resize', invert_depth: False, keep_model_size: False, da3_model: dict, images: torch.float32 [1,680,1208,3] cpu [05:32] [ComfyUI] [VRAM] Before DepthAnythingV3_MultiView: 386 / 7703 MB [05:32] [ComfyUI] [SubprocessWorker] call_method: nodes.nodes_multiview.DepthAnythingV3_MultiView.EXECUTE_NORMALIZED [05:32] [ComfyUI] [SubprocessWorker] acquired lock, ensuring started... [05:32] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [05:32] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ok [05:32] [ComfyUI] [SubprocessWorker] worker started/confirmed [05:32] [ComfyUI] [comfy-env] PRE-SERIALIZE 'images' shape: torch.Size([1, 680, 1208, 3]) [05:32] [ComfyUI] [SubprocessWorker] serializing kwargs to shm... [05:32] [ComfyUI] [SubprocessWorker] created 0 shm blocks [05:32] [ComfyUI] [SubprocessWorker] sending request via socket... [05:32] [ComfyUI] [SubprocessWorker] call_id=13 sending call_method [05:32] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [05:32] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[1, 680, 1208, 3] [05:32] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([1, 680, 1208, 3]) [05:33] [ComfyUI] [SubprocessWorker] callback 'request_vram_budget' from call_id=13 [05:33] [ComfyUI] [comfy-env] VRAM request: 0.82GB, free before eviction: 7.67GB [05:33] [ComfyUI] [comfy-env] VRAM after eviction: 7.67GB [05:33] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +482MB (now 482MB) reserved=512MB peak=482MB [05:34] Frame 3 saved (frame_002.jpg, t=5.0s) [05:34] [capture-periodic] freeze=1ms shot=142ms unfreeze=5ms saved=True [05:34] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +461MB (now 943MB) reserved=2848MB peak=1240MB [05:35] [ComfyUI] [SubprocessWorker] got response: ok [05:35] [ComfyUI] [comfy-env] Created 1 model patchers: ['DepthAnything3Net_1'] [05:35] [ComfyUI] [comfy-env] DepthAnythingV3_MultiView done (2.70s): result: NodeOutput [05:35] [ComfyUI] [VRAM] After DepthAnythingV3_MultiView: 3340 / 7703 MB [05:35] [ComfyUI] [VRAM] Loaded: SubprocessModel (783 MB) [05:35] [ComfyUI] [comfy-env] Running DepthAnythingV3_MultiView: normalization_mode: 'Raw', resize_method: 'resize', invert_depth: False, keep_model_size: False, da3_model: dict, images: torch.float32 [1,680,1208,3] cpu [05:35] [ComfyUI] [VRAM] Before DepthAnythingV3_MultiView: 3340 / 7703 MB [05:35] [ComfyUI] [VRAM] Loaded: SubprocessModel (783 MB) [05:35] [ComfyUI] [SubprocessWorker] call_method: nodes.nodes_multiview.DepthAnythingV3_MultiView.EXECUTE_NORMALIZED [05:35] [ComfyUI] [SubprocessWorker] acquired lock, ensuring started... [05:35] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [05:35] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ok [05:35] [ComfyUI] [SubprocessWorker] worker started/confirmed [05:35] [ComfyUI] [comfy-env] PRE-SERIALIZE 'images' shape: torch.Size([1, 680, 1208, 3]) [05:35] [capture-loop] iter=50 t=6.1s state={'complete': False, 'error': None, 'executedCount': 1, 'wsState': 1} eval_ms=2 ram=6.6/31.3GB (21%) swap=0.0/32.0GB (0%) cpu=14% browser_rss=0.88GB [05:35] [ComfyUI] [SubprocessWorker] serializing kwargs to shm... [05:35] [ComfyUI] [SubprocessWorker] created 0 shm blocks [05:35] [ComfyUI] [SubprocessWorker] sending request via socket... [05:35] [ComfyUI] [SubprocessWorker] call_id=14 sending call_method [05:35] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [05:35] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[1, 680, 1208, 3] [05:35] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([1, 680, 1208, 3]) [05:35] [ComfyUI] [SubprocessWorker] callback 'request_vram_budget' from call_id=14 [05:35] [ComfyUI] [comfy-env] VRAM request: 0.82GB, free before eviction: 4.58GB [05:35] [ComfyUI] [comfy-env] VRAM after eviction: 4.58GB [05:35] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +296MB (now 1240MB) reserved=1408MB peak=1240MB [05:36] [ComfyUI] [SubprocessWorker] got response: ok [05:36] [ComfyUI] [comfy-env] DepthAnythingV3_MultiView done (1.05s): result: NodeOutput [05:36] [ComfyUI] [VRAM] After DepthAnythingV3_MultiView: 3308 / 7703 MB [05:36] [ComfyUI] [VRAM] Loaded: SubprocessModel (783 MB) [05:36] [ComfyUI] [comfy-env] Running DA3_MultiViewPointCloud: extrinsics: '{\n "extrinsics": [\n {\n "image_0": [\n ..., intrinsics: '{\n "intrinsics": [\n {\n "image_0": [\n ..., confidence_threshold: 0.3, downsample: 4, use_icp: False, allow_around_1: False, filter_outliers: True, outlier_percentage: 5.0, depths: torch.float32 [1,680,1208,3] cpu, images: torch.float32 [1,680,1208,3] cpu, confidence: torch.float32 [1,680,1208,3] cpu, sky_mask: torch.float32 [1,680,1208] cpu [05:36] [ComfyUI] [VRAM] Before DA3_MultiViewPointCloud: 3308 / 7703 MB [05:36] [ComfyUI] [VRAM] Loaded: SubprocessModel (783 MB) [05:36] [ComfyUI] [SubprocessWorker] call_method: nodes.nodes_multiview.DA3_MultiViewPointCloud.EXECUTE_NORMALIZED [05:36] [ComfyUI] [SubprocessWorker] acquired lock, ensuring started... [05:36] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [05:36] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ok [05:36] [ComfyUI] [SubprocessWorker] worker started/confirmed [05:36] [ComfyUI] [comfy-env] PRE-SERIALIZE 'depths' shape: torch.Size([1, 680, 1208, 3]) [05:36] [ComfyUI] [comfy-env] PRE-SERIALIZE 'images' shape: torch.Size([1, 680, 1208, 3]) [05:36] [ComfyUI] [comfy-env] PRE-SERIALIZE 'confidence' shape: torch.Size([1, 680, 1208, 3]) [05:36] [ComfyUI] [comfy-env] PRE-SERIALIZE 'sky_mask' shape: torch.Size([1, 680, 1208]) [05:36] [ComfyUI] [SubprocessWorker] serializing kwargs to shm... [05:36] [ComfyUI] [SubprocessWorker] created 0 shm blocks [05:36] [ComfyUI] [SubprocessWorker] sending request via socket... [05:36] [ComfyUI] [SubprocessWorker] call_id=15 sending call_method [05:36] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [05:36] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[1, 680, 1208, 3] [05:36] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([1, 680, 1208, 3]) [05:36] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [05:36] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[1, 680, 1208, 3] [05:36] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([1, 680, 1208, 3]) [05:36] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [05:36] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[1, 680, 1208, 3] [05:36] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([1, 680, 1208, 3]) [05:36] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [05:36] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[1, 680, 1208] [05:36] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([1, 680, 1208]) [05:36] Frame 4 saved (frame_003.jpg, t=7.4s) [05:36] [capture-periodic] freeze=1ms shot=139ms unfreeze=5ms saved=True [05:36] [ComfyUI] [SubprocessWorker] callback 'report_progress' from call_id=15 [05:36] [ComfyUI] [SubprocessWorker] got response: ok [05:37] [ComfyUI] [ComfyUI-DepthAnythingV3] python: /workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/bin/python [05:37] [ComfyUI] [comfy-env] DA3_MultiViewPointCloud done (0.81s): result: NodeOutput [05:37] [ComfyUI] [ComfyUI-DepthAnythingV3] sys_path sent to worker: ['/workspaces/DepthAnythingV3-0307/ComfyUI', '/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3', '/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/lib/python3.11/site-packages', '/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3'] [05:37] [ComfyUI] [VRAM] After DA3_MultiViewPointCloud: 3308 / 7703 MB [05:37] [ComfyUI] [SubprocessWorker] is_pixi=True, python=/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/bin/python [05:37] [ComfyUI] [VRAM] Loaded: SubprocessModel (783 MB) [05:37] [ComfyUI] [SubprocessWorker] launching cmd=['/root/.pixi/bin/pixi', 'run', '--as-is', '--manifest-path', '/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/pixi.toml', '-e', 'depthanythingv3-nodes', 'python', '/tmp/comfyui_pvenv_4rmut23q/persistent_worker.py', 'abstract://comfy_worker_13ce4a65fabd']... [05:37] [ComfyUI] [comfy-env] Running DA3_MultiViewPointCloud: extrinsics: '{\n "extrinsics": [\n {\n "image_0": [\n ..., intrinsics: '{\n "intrinsics": [\n {\n "image_0": [\n ..., confidence_threshold: 0.3, downsample: 4, use_icp: False, allow_around_1: False, filter_outliers: True, outlier_percentage: 5.0, depths: torch.float32 [1,680,1208,3] cpu, images: torch.float32 [1,680,1208,3] cpu, confidence: torch.float32 [1,680,1208,3] cpu, sky_mask: torch.float32 [1,680,1208] cpu [05:37] [ComfyUI] [SubprocessWorker] socket_addr=abstract://comfy_worker_13ce4a65fabd (abstract, no filesystem path) [05:37] [ComfyUI] [VRAM] Before DA3_MultiViewPointCloud: 3308 / 7703 MB [05:37] [ComfyUI] [VRAM] Loaded: SubprocessModel (783 MB) [05:37] [ComfyUI] [SubprocessWorker] call_method: nodes.nodes_multiview.DA3_MultiViewPointCloud.EXECUTE_NORMALIZED [05:37] [ComfyUI] [SubprocessWorker] acquired lock, ensuring started... [05:37] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [05:37] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: failed [05:37] [ComfyUI] [ComfyUI-DepthAnythingV3] Socket unhealthy, restarting worker... [05:37] [ComfyUI] [comfy-env] Invalidated 1 stale model patchers (will be cleaned up during next unload) [05:37] [ComfyUI] Traceback (most recent call last): [05:37] [ComfyUI] File "/tmp/comfyui_pvenv_4rmut23q/persistent_worker.py", line 1807, in [05:37] [ComfyUI] main() [05:37] [ComfyUI] File "/tmp/comfyui_pvenv_4rmut23q/persistent_worker.py", line 1667, in main [05:37] [ComfyUI] transport.send({"status": "pong", "call_id": _current_call_id}) [05:37] [ComfyUI] File "/tmp/comfyui_pvenv_4rmut23q/persistent_worker.py", line 1075, in send [05:37] [ComfyUI] self._sock.sendall(msg) [05:37] [ComfyUI] BrokenPipeError: [Errno 32] Broken pipe [05:39] Frame 5 saved (frame_004.jpg, t=9.7s) [05:39] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] WARNING: WARNING: You need pytorch with cu130 or higher to use optimized CUDA operations. [05:39] [capture-periodic] freeze=1ms shot=143ms unfreeze=4ms saved=True [05:39] [ComfyUI] [SubprocessWorker] worker started/confirmed [05:39] [ComfyUI] [comfy-env] PRE-SERIALIZE 'depths' shape: torch.Size([1, 680, 1208, 3]) [05:39] [ComfyUI] [comfy-env] PRE-SERIALIZE 'images' shape: torch.Size([1, 680, 1208, 3]) [05:39] [ComfyUI] [comfy-env] PRE-SERIALIZE 'confidence' shape: torch.Size([1, 680, 1208, 3]) [05:39] [ComfyUI] [comfy-env] PRE-SERIALIZE 'sky_mask' shape: torch.Size([1, 680, 1208]) [05:39] [ComfyUI] [SubprocessWorker] serializing kwargs to shm... [05:39] [ComfyUI] [SubprocessWorker] created 0 shm blocks [05:39] [ComfyUI] [SubprocessWorker] sending request via socket... [05:39] [ComfyUI] [SubprocessWorker] call_id=16 sending call_method [05:39] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [05:39] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[1, 680, 1208, 3] [05:39] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([1, 680, 1208, 3]) [05:39] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [05:39] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[1, 680, 1208, 3] [05:39] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([1, 680, 1208, 3]) [05:39] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [05:39] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[1, 680, 1208, 3] [05:39] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([1, 680, 1208, 3]) [05:39] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [05:39] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[1, 680, 1208] [05:39] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([1, 680, 1208]) [05:40] [ComfyUI] [SubprocessWorker] callback 'report_progress' from call_id=16 [05:40] [ComfyUI] [SubprocessWorker] got response: ok [05:40] [ComfyUI] [comfy-env] [VRAM] Before offload 'DepthAnything3Net_1' (783 MB) -> cpu: 386 / 7703 MB [05:40] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [05:40] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ok [05:40] [ComfyUI] [SubprocessWorker] call_id=17 sending model_to_device [05:40] [ComfyUI] Model 'DepthAnything3Net_1' gone from worker (restarted?), treating as offloaded [05:40] [ComfyUI] [comfy-env] [VRAM] After offload 'DepthAnything3Net_1' (783 MB): 386 / 7703 MB [05:40] [ComfyUI] [comfy-env] DA3_MultiViewPointCloud done (3.38s): result: NodeOutput [05:40] [ComfyUI] [VRAM] After DA3_MultiViewPointCloud: 386 / 7703 MB [05:40] [ComfyUI] [comfy-env] Running DA3_SavePointCloud: filename_prefix: 'pointcloud', pointcloud: list[1] [05:40] [ComfyUI] [VRAM] Before DA3_SavePointCloud: 386 / 7703 MB [05:40] [ComfyUI] [SubprocessWorker] call_method: nodes.nodes_3d.DA3_SavePointCloud.EXECUTE_NORMALIZED [05:40] [ComfyUI] [SubprocessWorker] acquired lock, ensuring started... [05:40] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [05:40] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ok [05:40] [ComfyUI] [SubprocessWorker] worker started/confirmed [05:40] [ComfyUI] [SubprocessWorker] serializing kwargs to shm... [05:40] [ComfyUI] [SubprocessWorker] created 0 shm blocks [05:40] [ComfyUI] [SubprocessWorker] sending request via socket... [05:40] [ComfyUI] [SubprocessWorker] call_id=18 sending call_method [05:40] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [05:40] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[20447, 3] [05:40] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([20447, 3]) [05:40] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [05:40] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[20447, 3] [05:40] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([20447, 3]) [05:40] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [05:40] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[20447] [05:40] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([20447]) [05:40] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [05:41] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[20447] [05:41] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([20447]) [05:41] [capture-loop] iter=100 t=11.8s state={'complete': False, 'error': None, 'executedCount': 1, 'wsState': 1} eval_ms=2 ram=5.8/31.3GB (19%) swap=0.0/32.0GB (0%) cpu=12% browser_rss=0.89GB [05:41] [ComfyUI] [SubprocessWorker] got response: ok [05:41] [ComfyUI] [comfy-env] DA3_SavePointCloud done (0.92s): result: NodeOutput [05:41] [ComfyUI] [VRAM] After DA3_SavePointCloud: 386 / 7703 MB [05:41] [ComfyUI] [comfy-env] Running DA3_SavePointCloud: filename_prefix: 'pointcloud', pointcloud: list[1] [05:41] [ComfyUI] [VRAM] Before DA3_SavePointCloud: 386 / 7703 MB [05:41] [ComfyUI] [SubprocessWorker] call_method: nodes.nodes_3d.DA3_SavePointCloud.EXECUTE_NORMALIZED [05:41] [ComfyUI] [SubprocessWorker] acquired lock, ensuring started... [05:41] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [05:41] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ok [05:41] [ComfyUI] [SubprocessWorker] worker started/confirmed [05:41] [ComfyUI] [SubprocessWorker] serializing kwargs to shm... [05:41] Frame 6 saved (frame_005.jpg, t=12.0s) [05:41] [ComfyUI] [SubprocessWorker] created 0 shm blocks [05:41] [capture-periodic] freeze=2ms shot=139ms unfreeze=5ms saved=True [05:41] [ComfyUI] [SubprocessWorker] sending request via socket... [05:41] [ComfyUI] [SubprocessWorker] call_id=19 sending call_method [05:41] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [05:41] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[18819, 3] [05:41] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([18819, 3]) [05:41] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [05:41] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[18819, 3] [05:41] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([18819, 3]) [05:41] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [05:41] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[18819] [05:41] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([18819]) [05:41] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [05:41] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[18819] [05:41] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([18819]) [05:42] [ComfyUI] [SubprocessWorker] got response: ok [05:42] [ComfyUI] [comfy-env] DA3_SavePointCloud done (0.84s): result: NodeOutput [05:42] [ComfyUI] [VRAM] After DA3_SavePointCloud: 386 / 7703 MB [05:42] Node executed (2 total), capturing... [05:42] [ComfyUI] [comfy-env] Running DA3_PreviewPointCloud: file_path: '/workspaces/DepthAnythingV3-0307/ComfyUI/output/pointclo..., color_mode: 'RGB' [05:42] [ComfyUI] [VRAM] Before DA3_PreviewPointCloud: 386 / 7703 MB [05:42] [ComfyUI] [SubprocessWorker] call_method: nodes.preview_nodes.DA3_PreviewPointCloud.EXECUTE_NORMALIZED [05:42] [ComfyUI] [SubprocessWorker] acquired lock, ensuring started... [05:42] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [05:42] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ok [05:42] [ComfyUI] [SubprocessWorker] worker started/confirmed [05:42] [ComfyUI] [SubprocessWorker] serializing kwargs to shm... [05:42] [ComfyUI] [SubprocessWorker] created 0 shm blocks [05:42] [ComfyUI] [SubprocessWorker] sending request via socket... [05:42] [ComfyUI] [SubprocessWorker] call_id=20 sending call_method [05:42] [ComfyUI] [SubprocessWorker] got response: ok [05:42] [ComfyUI] [comfy-env] DA3_PreviewPointCloud done (0.30s): result: NodeOutput [05:42] [ComfyUI] [VRAM] After DA3_PreviewPointCloud: 386 / 7703 MB [05:42] [ComfyUI] [comfy-env] Running DA3_PreviewPointCloud: file_path: '/workspaces/DepthAnythingV3-0307/ComfyUI/output/pointclo..., color_mode: 'RGB' [05:42] Frame 7 saved (frame_006.jpg, t=13.0s) [05:42] [ComfyUI] [VRAM] Before DA3_PreviewPointCloud: 386 / 7703 MB [05:42] [capture-node] freeze=2ms shot=137ms unfreeze=5ms saved=True [05:42] [ComfyUI] [SubprocessWorker] call_method: nodes.preview_nodes.DA3_PreviewPointCloud.EXECUTE_NORMALIZED [05:42] [ComfyUI] [SubprocessWorker] acquired lock, ensuring started... [05:42] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [05:42] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ok [05:42] [ComfyUI] [SubprocessWorker] worker started/confirmed [05:42] [ComfyUI] [SubprocessWorker] serializing kwargs to shm... [05:42] [ComfyUI] [SubprocessWorker] created 0 shm blocks [05:42] [ComfyUI] [SubprocessWorker] sending request via socket... [05:42] [ComfyUI] [SubprocessWorker] call_id=21 sending call_method [05:42] [ComfyUI] [SubprocessWorker] got response: ok [05:42] Node executed (3 total), capturing... [05:42] [ComfyUI] [comfy-env] DA3_PreviewPointCloud done (0.31s): result: NodeOutput [05:42] [ComfyUI] [VRAM] After DA3_PreviewPointCloud: 386 / 7703 MB [05:42] [ComfyUI] Prompt executed in 13.19 seconds [05:43] Frame 8 saved (frame_007.jpg, t=13.6s) [05:43] [capture-node] freeze=2ms shot=230ms unfreeze=5ms saved=True [05:43] Node executed (4 total), capturing... [05:43] Frame 9 saved (frame_008.jpg, t=14.2s) [05:43] [capture-node] freeze=2ms shot=139ms unfreeze=0ms saved=True [05:43] Execution complete (t=14.2s) [05:43] Captured 9 unique frames over 14.38s [05:43] [debug] iframes: 1 [05:43] [debug] iframe[0]: src=http://127.0.0.1:41886/extensions/ComfyUI-DepthAnythingV3/viewer_vtk.html?v=1778, canvas=486x394, rafFrozen=True [05:43] [timing] trigger_3d_previews: 0.0s [05:46] Saved high-quality screenshot: advanced_3d_multiview_executed.png [05:46] [timing] final_screenshot: 0.3s [05:46] Capturing execution frames: bas_relief.json [05:52] Frame 1 saved (frame_000.jpg, t=0.0s) [05:52] Validating workflow... [05:52] [ComfyUI] === /validate endpoint called === [05:52] [ComfyUI] Received JSON with keys: ['prompt'] [05:52] Queuing workflow for execution... [05:52] [ComfyUI] Prompt has 6 nodes [05:52] [ComfyUI] Node 2: LoadImage [05:52] [ComfyUI] Inputs: {"image": "bas_relief.png"} [05:52] [ComfyUI] Node 3: DownloadAndLoadDepthAnythingV3Model [05:52] [ComfyUI] Inputs: {"model": "da3_large.safetensors", "precision": "auto", "attention": "auto"} [05:52] [ComfyUI] Node 21: DepthAnything_V3 [05:52] [ComfyUI] Inputs: {"normalization_mode": "Raw", "resize_method": "resize", "invert_depth": false, "keep_model_size": true, "da3_model": ["3", 0], "images": ["2", 0]} [05:52] [ComfyUI] Node 27: PreviewImage [05:52] [ComfyUI] Inputs: {"images": ["21", 2]} [05:52] [ComfyUI] Node 30: Preview3D [05:52] [ComfyUI] Inputs: {"model_file": ["32", 0], "image": ""} [05:52] [ComfyUI] Node 32: DA3_ToMesh [05:52] [ComfyUI] Inputs: {"confidence_threshold": 0.1, "depth_edge_threshold": 0.1, "downsample": 1, "target_faces": 100000, "filename_prefix": "mesh", "allow_around_1": true, "use_draco_compression": false, "depth_raw": ["21 [05:52] [ComfyUI] Generated prompt_id: 2b75abcf-886c-4288-bfcb-d6ea6136547d [05:52] [ComfyUI] Calling execution.validate_prompt()... [05:52] [ComfyUI] Validation result: valid=True [05:52] [ComfyUI] === Validation PASSED === [05:52] [ComfyUI] got prompt [05:52] [ComfyUI] [comfy-env] Running DownloadAndLoadDepthAnythingV3Model: model: 'da3_large.safetensors', precision: 'auto', attention: 'auto' [05:52] [ComfyUI] [VRAM] Before DownloadAndLoadDepthAnythingV3Model: 386 / 7703 MB [05:52] [ComfyUI] [SubprocessWorker] call_method: nodes.load_model.DownloadAndLoadDepthAnythingV3Model.EXECUTE_NORMALIZED [05:52] [ComfyUI] [SubprocessWorker] acquired lock, ensuring started... [05:52] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [05:52] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ok [05:52] [ComfyUI] [SubprocessWorker] worker started/confirmed [05:52] [ComfyUI] [SubprocessWorker] serializing kwargs to shm... [05:52] [ComfyUI] [SubprocessWorker] created 0 shm blocks [05:52] [ComfyUI] [SubprocessWorker] sending request via socket... [05:52] [ComfyUI] [SubprocessWorker] call_id=22 sending call_method [05:52] [ComfyUI] [SubprocessWorker] got response: ok [05:52] [ComfyUI] [comfy-env] DownloadAndLoadDepthAnythingV3Model done (0.11s): result: NodeOutput [05:52] [ComfyUI] [VRAM] After DownloadAndLoadDepthAnythingV3Model: 386 / 7703 MB [05:52] [ComfyUI] [comfy-env] Running DepthAnything_V3: normalization_mode: 'Raw', resize_method: 'resize', invert_depth: False, keep_model_size: True, da3_model: dict, images: torch.float32 [1,2000,1336,3] cpu [05:52] [ComfyUI] [VRAM] Before DepthAnything_V3: 386 / 7703 MB [05:52] [ComfyUI] [SubprocessWorker] call_method: nodes.nodes_inference.DepthAnything_V3.EXECUTE_NORMALIZED [05:52] [ComfyUI] [SubprocessWorker] acquired lock, ensuring started... [05:52] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [05:52] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ok [05:52] [ComfyUI] [SubprocessWorker] worker started/confirmed [05:52] [ComfyUI] [comfy-env] PRE-SERIALIZE 'images' shape: torch.Size([1, 2000, 1336, 3]) [05:52] [ComfyUI] [SubprocessWorker] serializing kwargs to shm... [05:52] [ComfyUI] [SubprocessWorker] created 0 shm blocks [05:52] [ComfyUI] [SubprocessWorker] sending request via socket... [05:52] [ComfyUI] [SubprocessWorker] call_id=23 sending call_method [05:52] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [05:52] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[1, 2000, 1336, 3] [05:52] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([1, 2000, 1336, 3]) [05:53] [ComfyUI] [SubprocessWorker] callback 'request_vram_budget' from call_id=23 [05:53] [ComfyUI] [comfy-env] VRAM request: 0.82GB, free before eviction: 7.67GB [05:53] [ComfyUI] [comfy-env] VRAM after eviction: 7.67GB [05:54] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +731MB (now 731MB) reserved=736MB peak=731MB [05:54] Frame 2 saved (frame_001.jpg, t=2.4s) [05:54] [capture-periodic] freeze=2ms shot=127ms unfreeze=3ms saved=True [05:55] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +2749MB (now 3480MB) reserved=6240MB peak=3480MB [05:55] [ComfyUI] [SubprocessWorker] callback 'report_progress' from call_id=23 [05:56] [ComfyUI] [SubprocessWorker] got response: ok [05:56] [ComfyUI] [comfy-env] Created 1 model patchers: ['DepthAnything3Net_1'] [05:56] [ComfyUI] [comfy-env] DepthAnything_V3 done (3.37s): result: NodeOutput [05:56] [ComfyUI] [VRAM] After DepthAnything_V3: 7500 / 7703 MB [05:56] [ComfyUI] [VRAM] Loaded: SubprocessModel (783 MB) [05:56] [ComfyUI] [comfy-env] Running DA3_ToMesh: confidence_threshold: 0.1, depth_edge_threshold: 0.1, downsample: 1, target_faces: 100000, filename_prefix: 'mesh', allow_around_1: True, use_draco_compression: False, depth_raw: torch.float32 [1,2002,1330,3] cpu, confidence: torch.float32 [1,2002,1330,3] cpu, intrinsics: '{\n "intrinsics": [\n {\n "image_0": [\n ..., sky_mask: torch.float32 [1,2002,1330] cpu, source_image: torch.float32 [1,2002,1330,3] cpu [05:56] [ComfyUI] [VRAM] Before DA3_ToMesh: 7500 / 7703 MB [05:56] [ComfyUI] [VRAM] Loaded: SubprocessModel (783 MB) [05:56] [ComfyUI] [SubprocessWorker] call_method: nodes.nodes_3d.DA3_ToMesh.EXECUTE_NORMALIZED [05:56] [ComfyUI] [SubprocessWorker] acquired lock, ensuring started... [05:56] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [05:56] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ok [05:56] [ComfyUI] [SubprocessWorker] worker started/confirmed [05:56] [ComfyUI] [comfy-env] PRE-SERIALIZE 'depth_raw' shape: torch.Size([1, 2002, 1330, 3]) [05:56] [ComfyUI] [comfy-env] PRE-SERIALIZE 'confidence' shape: torch.Size([1, 2002, 1330, 3]) [05:56] [ComfyUI] [comfy-env] PRE-SERIALIZE 'sky_mask' shape: torch.Size([1, 2002, 1330]) [05:56] [ComfyUI] [comfy-env] PRE-SERIALIZE 'source_image' shape: torch.Size([1, 2002, 1330, 3]) [05:56] [ComfyUI] [SubprocessWorker] serializing kwargs to shm... [05:56] [ComfyUI] [SubprocessWorker] created 0 shm blocks [05:56] [ComfyUI] [SubprocessWorker] sending request via socket... [05:56] [ComfyUI] [SubprocessWorker] call_id=24 sending call_method [05:56] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [05:56] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[1, 2002, 1330, 3] [05:56] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([1, 2002, 1330, 3]) [05:56] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [05:56] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[1, 2002, 1330, 3] [05:56] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([1, 2002, 1330, 3]) [05:56] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [05:56] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[1, 2002, 1330] [05:56] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([1, 2002, 1330]) [05:56] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [05:56] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[1, 2002, 1330, 3] [05:56] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([1, 2002, 1330, 3]) [05:56] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] -2688MB (now 793MB) reserved=7008MB peak=3480MB [05:56] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] WARNING: pyfqmr not installed, skipping decimation. Install with: pip install pyfqmr [05:56] Frame 3 saved (frame_002.jpg, t=4.6s) [05:56] [capture-periodic] freeze=1ms shot=163ms unfreeze=3ms saved=True [05:57] [capture-loop] iter=50 t=5.8s state={'complete': False, 'error': None, 'executedCount': 0, 'wsState': 1} eval_ms=2 ram=7.9/31.3GB (25%) swap=0.0/32.0GB (0%) cpu=11% browser_rss=1.06GB [05:59] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] WARNING: pygltflib is required for material post-processing. Install with: pip install pygltflib. Mesh may appear dark or single-sided. [05:59] [ComfyUI] [SubprocessWorker] got response: ok [05:59] [ComfyUI] [comfy-env] DA3_ToMesh done (3.49s): result: NodeOutput [05:59] [ComfyUI] [VRAM] After DA3_ToMesh: 7568 / 7703 MB [05:59] [ComfyUI] [VRAM] Loaded: SubprocessModel (783 MB) [05:59] Node executed (2 total), capturing... [05:59] [ComfyUI] Prompt executed in 7.41 seconds [06:00] Frame 4 saved (frame_003.jpg, t=7.7s) [06:00] [capture-node] freeze=25ms shot=149ms unfreeze=2ms saved=True [06:00] Node executed (3 total), capturing... [06:03] Frame 5 saved (frame_004.jpg, t=8.3s) [06:03] [capture-node] freeze=56ms shot=3246ms unfreeze=0ms saved=True [06:03] Execution complete (t=8.3s) [06:04] [3d-wait] Phase 1: Waiting for 'Loading 3D Model' text to disappear... [07:04] [3d-wait] Phase 1 exception after 60.0s: Page.wait_for_function: Timeout 60000ms exceeded. [07:04] [3d-wait] Phase 2: iframes=0 [07:04] [3d-wait] Total 3D wait: 60.0s [07:04] Captured 5 unique frames over 71.95s [07:04] [debug] iframes: 0 [07:04] [timing] trigger_3d_previews: 0.0s [07:06] Saved high-quality screenshot: bas_relief_executed.png [07:06] [timing] final_screenshot: 0.8s [07:06] [ComfyUI] [comfy-env] [VRAM] Before offload 'DepthAnything3Net_1' (783 MB) -> cpu: 7568 / 7703 MB [07:06] Capturing execution frames: da3_streaming.json [07:06] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [07:06] [ComfyUI] [ComfyUI-DepthAnythingV3] python: /workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/bin/python [07:06] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: failed [07:06] [ComfyUI] [ComfyUI-DepthAnythingV3] Socket unhealthy, restarting worker... [07:06] [ComfyUI] [ComfyUI-DepthAnythingV3] sys_path sent to worker: ['/workspaces/DepthAnythingV3-0307/ComfyUI', '/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3', '/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/lib/python3.11/site-packages', '/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3'] [07:06] [ComfyUI] [comfy-env] Invalidated 1 stale model patchers (will be cleaned up during next unload) [07:06] [ComfyUI] [SubprocessWorker] is_pixi=True, python=/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/bin/python [07:06] [ComfyUI] [SubprocessWorker] launching cmd=['/root/.pixi/bin/pixi', 'run', '--as-is', '--manifest-path', '/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/pixi.toml', '-e', 'depthanythingv3-nodes', 'python', '/tmp/comfyui_pvenv_4rmut23q/persistent_worker.py', 'abstract://comfy_worker_d5bb162f38f7']... [07:06] [ComfyUI] [SubprocessWorker] socket_addr=abstract://comfy_worker_d5bb162f38f7 (abstract, no filesystem path) [07:09] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] WARNING: WARNING: You need pytorch with cu130 or higher to use optimized CUDA operations. [07:10] [ComfyUI] [SubprocessWorker] call_id=25 sending model_to_device [07:10] [ComfyUI] Model 'DepthAnything3Net_1' gone from worker (restarted?), treating as offloaded [07:10] [ComfyUI] [comfy-env] [VRAM] After offload 'DepthAnything3Net_1' (783 MB): 386 / 7703 MB [07:12] Frame 1 saved (frame_000.jpg, t=0.0s) [07:12] Validating workflow... [07:12] [ComfyUI] === /validate endpoint called === [07:12] [ComfyUI] Received JSON with keys: ['prompt'] [07:12] Queuing workflow for execution... [07:12] [ComfyUI] Prompt has 5 nodes [07:12] [ComfyUI] Node 2: DownloadAndLoadDepthAnythingV3Model [07:12] [ComfyUI] Inputs: {"model": "da3_large.safetensors", "precision": "auto", "attention": "auto"} [07:12] [ComfyUI] Node 3: LoadVideo [07:12] [ComfyUI] Inputs: {"file": "robot_unitree.mp4", "video-preview": ""} [07:12] [ComfyUI] Node 4: SaveVideo [07:12] [ComfyUI] Inputs: {"filename_prefix": "video/preview_da3", "format": "auto", "codec": "auto", "video": ["6", 0]} [07:12] [ComfyUI] Node 5: LoadSALADModel [07:12] [ComfyUI] Inputs: {} [07:12] [ComfyUI] Node 6: DepthAnythingV3_Streaming [07:12] [ComfyUI] Inputs: {"normalization_mode": "Standard", "chunk_size": 30, "overlap": 8, "align_lib": "auto", "align_method": "sim3", "resize_method": "resize", "invert_depth": true, "save_pointcloud": false, "sample_ratio [07:12] [ComfyUI] Generated prompt_id: b9bd0d88-ae31-4128-a017-71456281ecea [07:12] [ComfyUI] Calling execution.validate_prompt()... [07:12] [ComfyUI] Validation result: valid=True [07:12] [ComfyUI] === Validation PASSED === [07:12] [ComfyUI] got prompt [07:12] [ComfyUI] [comfy-env] Running LoadSALADModel: [07:12] [ComfyUI] [VRAM] Before LoadSALADModel: 386 / 7703 MB [07:12] [ComfyUI] [SubprocessWorker] call_method: nodes.load_model.LoadSALADModel.EXECUTE_NORMALIZED [07:12] [ComfyUI] [SubprocessWorker] acquired lock, ensuring started... [07:12] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [07:12] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ok [07:12] [ComfyUI] [SubprocessWorker] worker started/confirmed [07:12] [ComfyUI] [SubprocessWorker] sending request via socket... [07:12] [ComfyUI] [SubprocessWorker] call_id=26 sending call_method [07:13] [ComfyUI] Warning: You are sending unauthenticated requests to the HF Hub. Please set a HF_TOKEN to enable higher rate limits and faster downloads. [07:13] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] WARNING: Warning: You are sending unauthenticated requests to the HF Hub. Please set a HF_TOKEN to enable higher rate limits and faster downloads. [07:13] [ComfyUI] [07:15] Frame 2 saved (frame_001.jpg, t=2.3s) [07:15] [capture-periodic] freeze=2ms shot=153ms unfreeze=4ms saved=True [07:17] Frame 3 saved (frame_002.jpg, t=4.6s) [07:17] [capture-periodic] freeze=2ms shot=112ms unfreeze=2ms saved=True [07:17] [ComfyUI] dino_salad.safetensors: 0%| | 0.00/352M [00:00 cpu: 386 / 7703 MB [07:18] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [07:18] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ok [07:18] [ComfyUI] [SubprocessWorker] call_id=27 sending model_to_device [07:18] [ComfyUI] Model 'DepthAnything3Net_1' gone from worker (restarted?), treating as offloaded [07:18] [ComfyUI] [comfy-env] [VRAM] After offload 'DepthAnything3Net_1' (783 MB): 386 / 7703 MB [07:18] [ComfyUI] [comfy-env] LoadSALADModel done (5.59s): result: NodeOutput [07:18] [ComfyUI] [VRAM] After LoadSALADModel: 386 / 7703 MB [07:18] [ComfyUI] [comfy-env] Running DownloadAndLoadDepthAnythingV3Model: model: 'da3_large.safetensors', precision: 'auto', attention: 'auto' [07:18] [ComfyUI] [VRAM] Before DownloadAndLoadDepthAnythingV3Model: 386 / 7703 MB [07:18] [ComfyUI] [SubprocessWorker] call_method: nodes.load_model.DownloadAndLoadDepthAnythingV3Model.EXECUTE_NORMALIZED [07:18] [ComfyUI] [SubprocessWorker] acquired lock, ensuring started... [07:18] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [07:18] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ok [07:18] [ComfyUI] [SubprocessWorker] worker started/confirmed [07:18] [ComfyUI] [SubprocessWorker] serializing kwargs to shm... [07:18] [ComfyUI] [SubprocessWorker] created 0 shm blocks [07:18] [ComfyUI] [SubprocessWorker] sending request via socket... [07:18] [ComfyUI] [SubprocessWorker] call_id=28 sending call_method [07:18] [ComfyUI] [SubprocessWorker] got response: ok [07:18] [ComfyUI] [comfy-env] DownloadAndLoadDepthAnythingV3Model done (0.10s): result: NodeOutput [07:18] [ComfyUI] [VRAM] After DownloadAndLoadDepthAnythingV3Model: 386 / 7703 MB [07:18] [ComfyUI] [comfy-env] Running DepthAnythingV3_Streaming: normalization_mode: 'Standard', chunk_size: 30, overlap: 8, align_lib: 'auto', align_method: 'sim3', resize_method: 'resize', invert_depth: True, save_pointcloud: False, sample_ratio: 0.015, conf_threshold_coef: 0.75, da3_model: dict, video: VideoFromFile, salad_model: dict [07:18] [ComfyUI] [VRAM] Before DepthAnythingV3_Streaming: 386 / 7703 MB [07:18] [ComfyUI] [SubprocessWorker] call_method: nodes.streaming.node.DepthAnythingV3_Streaming.EXECUTE_NORMALIZED [07:18] [ComfyUI] [SubprocessWorker] acquired lock, ensuring started... [07:18] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [07:18] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ok [07:18] [ComfyUI] [SubprocessWorker] worker started/confirmed [07:18] [ComfyUI] [SubprocessWorker] serializing kwargs to shm... [07:18] [ComfyUI] [SubprocessWorker] created 1 shm blocks [07:18] [ComfyUI] [SubprocessWorker] sending request via socket... [07:18] [ComfyUI] [SubprocessWorker] call_id=29 sending call_method [07:19] Frame 4 saved (frame_003.jpg, t=6.8s) [07:19] [capture-periodic] freeze=1ms shot=99ms unfreeze=1ms saved=True [07:21] [ComfyUI] [SubprocessWorker] callback 'request_vram_budget' from call_id=29 [07:21] [ComfyUI] [comfy-env] VRAM request: 1.17GB, free before eviction: 7.67GB [07:21] [ComfyUI] [comfy-env] VRAM after eviction: 7.67GB [07:21] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +315MB (now 315MB) reserved=320MB peak=315MB [07:22] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +804MB (now 1119MB) reserved=1120MB peak=1119MB [07:23] [capture-loop] iter=100 t=11.3s state={'complete': False, 'error': None, 'executedCount': 0, 'wsState': 1} eval_ms=2 ram=11.2/31.3GB (36%) swap=0.0/32.0GB (0%) cpu=17% browser_rss=1.06GB [07:25] [ComfyUI] [SubprocessWorker] callback 'report_progress' from call_id=29 [07:25] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +277MB (now 1396MB) reserved=1408MB peak=1396MB [07:26] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +1238MB (now 2635MB) reserved=4032MB peak=2635MB [07:27] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +350MB (now 2986MB) reserved=4480MB peak=3688MB [07:28] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +350MB (now 3337MB) reserved=4928MB peak=3688MB [07:29] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +350MB (now 3688MB) reserved=5280MB peak=3688MB [07:29] [capture-loop] iter=150 t=16.9s state={'complete': False, 'error': None, 'executedCount': 0, 'wsState': 1} eval_ms=3 ram=9.3/31.3GB (30%) swap=0.0/32.0GB (0%) cpu=15% browser_rss=1.05GB [07:30] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +1339MB (now 5027MB) reserved=5504MB peak=5027MB [07:31] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] -699MB (now 4329MB) reserved=4832MB peak=5027MB [07:34] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +206MB (now 4535MB) reserved=5760MB peak=5027MB [07:34] [capture-loop] iter=200 t=22.3s state={'complete': False, 'error': None, 'executedCount': 0, 'wsState': 1} eval_ms=2 ram=9.3/31.3GB (30%) swap=0.0/32.0GB (0%) cpu=15% browser_rss=1.05GB [07:35] [ComfyUI] [SubprocessWorker] callback 'report_progress' from call_id=29 [07:35] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] -2076MB (now 2460MB) reserved=3520MB peak=5027MB [07:36] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +526MB (now 2986MB) reserved=4160MB peak=5027MB [07:37] Frame 5 saved (frame_004.jpg, t=24.5s) [07:37] [capture-periodic] freeze=2ms shot=126ms unfreeze=3ms saved=True [07:37] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +350MB (now 3337MB) reserved=4576MB peak=5027MB [07:38] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +350MB (now 3688MB) reserved=4928MB peak=5027MB [07:40] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +568MB (now 4257MB) reserved=4544MB peak=5027MB [07:40] [capture-loop] iter=250 t=27.9s state={'complete': False, 'error': None, 'executedCount': 0, 'wsState': 1} eval_ms=2 ram=9.3/31.3GB (30%) swap=0.0/32.0GB (0%) cpu=15% browser_rss=0.96GB [07:43] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +206MB (now 4463MB) reserved=5472MB peak=5027MB [07:44] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] -2947MB (now 1517MB) reserved=7168MB peak=5027MB [07:44] [ComfyUI] [SubprocessWorker] callback 'report_progress' from call_id=29 [07:45] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +1118MB (now 2635MB) reserved=4064MB peak=5027MB [07:45] [capture-loop] iter=300 t=33.4s state={'complete': False, 'error': None, 'executedCount': 0, 'wsState': 1} eval_ms=2 ram=9.5/31.3GB (30%) swap=0.0/32.0GB (0%) cpu=16% browser_rss=0.96GB [07:46] Frame 6 saved (frame_005.jpg, t=33.4s) [07:46] [capture-periodic] freeze=2ms shot=142ms unfreeze=4ms saved=True [07:46] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +350MB (now 2986MB) reserved=4416MB peak=5027MB [07:47] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +350MB (now 3337MB) reserved=4832MB peak=5027MB [07:48] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +350MB (now 3688MB) reserved=5120MB peak=5027MB [07:49] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +568MB (now 4257MB) reserved=4544MB peak=5027MB [07:51] [capture-loop] iter=350 t=39.0s state={'complete': False, 'error': None, 'executedCount': 0, 'wsState': 1} eval_ms=2 ram=9.4/31.3GB (30%) swap=0.0/32.0GB (0%) cpu=13% browser_rss=0.96GB [07:53] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +206MB (now 4463MB) reserved=5472MB peak=5027MB [07:54] [ComfyUI] [SubprocessWorker] got response: error [07:54] [ComfyUI] !!! Exception during processing !!! Allocation on device 0 would exceed allowed memory. (out of memory) [07:54] [ComfyUI] Currently allocated : 4.92 GiB [07:54] [ComfyUI] Requested : 286.44 MiB [07:54] [ComfyUI] Device limit : 7.52 GiB [07:54] [ComfyUI] Free (according to CUDA): 11.19 MiB [07:54] [ComfyUI] PyTorch limit (set by user-supplied memory fraction) [07:54] [ComfyUI] : 17179869184.00 GiB [07:54] [ComfyUI] [07:54] [ComfyUI] Worker traceback: [07:54] [ComfyUI] Traceback (most recent call last): [07:54] [ComfyUI] File "/tmp/comfyui_pvenv_4rmut23q/persistent_worker.py", line 1762, in main [07:54] [ComfyUI] result = method(**inputs) [07:54] [ComfyUI] ^^^^^^^^^^^^^^^^ [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/ComfyUI/comfy_api/latest/_io.py", line 1833, in EXECUTE_NORMALIZED [07:54] [ComfyUI] to_return = cls.execute(*args, **kwargs) [07:54] [ComfyUI] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^ [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/streaming/node.py", line 249, in execute [07:54] [ComfyUI] result = pipeline.run(normalized_images, pbar=pbar, [07:54] [ComfyUI] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/streaming/pipeline.py", line 277, in run [07:54] [ComfyUI] result = self._process_chunk(normalized_images, start, end) [07:54] [ComfyUI] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/streaming/pipeline.py", line 330, in _process_chunk [07:54] [ComfyUI] output = self.model(chunk_input) [07:54] [ComfyUI] ^^^^^^^^^^^^^^^^^^^^^^^ [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1779, in _wrapped_call_impl [07:54] [ComfyUI] return self._call_impl(*args, **kwargs) [07:54] [ComfyUI] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1790, in _call_impl [07:54] [ComfyUI] return forward_call(*args, **kwargs) [07:54] [ComfyUI] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/depth_anything_v3/model.py", line 1577, in forward [07:54] [ComfyUI] output = self._process_depth_head(feats, H, W) [07:54] [ComfyUI] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/depth_anything_v3/model.py", line 1591, in _process_depth_head [07:54] [ComfyUI] return self.head(feats, H, W, patch_start_idx=0) [07:54] [ComfyUI] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1779, in _wrapped_call_impl [07:54] [ComfyUI] return self._call_impl(*args, **kwargs) [07:54] [ComfyUI] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1790, in _call_impl [07:54] [ComfyUI] return forward_call(*args, **kwargs) [07:54] [ComfyUI] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/depth_anything_v3/model.py", line 1379, in forward [07:54] [ComfyUI] self._forward_impl([feat[s0:s1] for feat in feats], H, W, patch_start_idx) [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/depth_anything_v3/model.py", line 1419, in _forward_impl [07:54] [ComfyUI] fused_main = self._add_pos_embed(fused_main, W, H) [07:54] [ComfyUI] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/depth_anything_v3/model.py", line 1477, in _add_pos_embed [07:54] [ComfyUI] pe = position_grid_to_embed(pe, x.shape[1]) * ratio [07:54] [ComfyUI] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~ [07:54] [ComfyUI] torch.OutOfMemoryError: Allocation on device 0 would exceed allowed memory. (out of memory) [07:54] [ComfyUI] Currently allocated : 4.92 GiB [07:54] [ComfyUI] Requested : 286.44 MiB [07:54] [ComfyUI] Device limit : 7.52 GiB [07:54] [ComfyUI] Free (according to CUDA): 11.19 MiB [07:54] [ComfyUI] PyTorch limit (set by user-supplied memory fraction) [07:54] [ComfyUI] : 17179869184.00 GiB [07:54] [ComfyUI] [07:54] [ComfyUI] Traceback (most recent call last): [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/ComfyUI/execution.py", line 535, in execute [07:54] [ComfyUI] output_data, output_ui, has_subgraph, has_pending_tasks = await get_output_data(prompt_id, unique_id, obj, input_data_all, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, v3_data=v3_data) [07:54] [ComfyUI] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/ComfyUI/execution.py", line 335, in get_output_data [07:54] [ComfyUI] return_values = await _async_map_node_over_list(prompt_id, unique_id, obj, input_data_all, obj.FUNCTION, allow_interrupt=True, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, v3_data=v3_data) [07:54] [ComfyUI] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/ComfyUI/execution.py", line 309, in _async_map_node_over_list [07:54] [ComfyUI] await process_inputs(input_dict, i) [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/ComfyUI/execution.py", line 297, in process_inputs [07:54] [ComfyUI] result = f(**inputs) [07:54] [ComfyUI] ^^^^^^^^^^^ [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/.venv/lib/python3.11/site-packages/comfy_env/isolation/metadata.py", line 536, in proxy [07:54] [ComfyUI] result = worker.call_method( [07:54] [ComfyUI] ^^^^^^^^^^^^^^^^^^^ [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/.venv/lib/python3.11/site-packages/comfy_env/isolation/workers/subprocess.py", line 674, in call_method [07:54] [ComfyUI] raise WorkerError( [07:54] [ComfyUI] comfy_env.isolation.workers.base.WorkerError: Allocation on device 0 would exceed allowed memory. (out of memory) [07:54] [ComfyUI] Currently allocated : 4.92 GiB [07:54] [ComfyUI] Requested : 286.44 MiB [07:54] [ComfyUI] Device limit : 7.52 GiB [07:54] [ComfyUI] Free (according to CUDA): 11.19 MiB [07:54] [ComfyUI] PyTorch limit (set by user-supplied memory fraction) [07:54] [ComfyUI] : 17179869184.00 GiB [07:54] [ComfyUI] [07:54] [ComfyUI] Worker traceback: [07:54] [ComfyUI] Traceback (most recent call last): [07:54] [ComfyUI] File "/tmp/comfyui_pvenv_4rmut23q/persistent_worker.py", line 1762, in main [07:54] [ComfyUI] result = method(**inputs) [07:54] [ComfyUI] ^^^^^^^^^^^^^^^^ [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/ComfyUI/comfy_api/latest/_io.py", line 1833, in EXECUTE_NORMALIZED [07:54] [ComfyUI] to_return = cls.execute(*args, **kwargs) [07:54] [ComfyUI] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^ [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/streaming/node.py", line 249, in execute [07:54] [ComfyUI] result = pipeline.run(normalized_images, pbar=pbar, [07:54] [ComfyUI] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/streaming/pipeline.py", line 277, in run [07:54] [ComfyUI] result = self._process_chunk(normalized_images, start, end) [07:54] [ComfyUI] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/streaming/pipeline.py", line 330, in _process_chunk [07:54] [ComfyUI] output = self.model(chunk_input) [07:54] [ComfyUI] ^^^^^^^^^^^^^^^^^^^^^^^ [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1779, in _wrapped_call_impl [07:54] [ComfyUI] return self._call_impl(*args, **kwargs) [07:54] [ComfyUI] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1790, in _call_impl [07:54] [ComfyUI] return forward_call(*args, **kwargs) [07:54] [ComfyUI] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/depth_anything_v3/model.py", line 1577, in forward [07:54] [ComfyUI] output = self._process_depth_head(feats, H, W) [07:54] [ComfyUI] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/depth_anything_v3/model.py", line 1591, in _process_depth_head [07:54] [ComfyUI] return self.head(feats, H, W, patch_start_idx=0) [07:54] [ComfyUI] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1779, in _wrapped_call_impl [07:54] [ComfyUI] return self._call_impl(*args, **kwargs) [07:54] [ComfyUI] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1790, in _call_impl [07:54] [ComfyUI] return forward_call(*args, **kwargs) [07:54] [ComfyUI] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/depth_anything_v3/model.py", line 1379, in forward [07:54] [ComfyUI] self._forward_impl([feat[s0:s1] for feat in feats], H, W, patch_start_idx) [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/depth_anything_v3/model.py", line 1419, in _forward_impl [07:54] [ComfyUI] fused_main = self._add_pos_embed(fused_main, W, H) [07:54] [ComfyUI] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ [07:54] [ComfyUI] File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/depth_anything_v3/model.py", line 1477, in _add_pos_embed [07:54] [ComfyUI] pe = position_grid_to_embed(pe, x.shape[1]) * ratio [07:54] [ComfyUI] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~ [07:54] [ComfyUI] torch.OutOfMemoryError: Allocation on device 0 would exceed allowed memory. (out of memory) [07:54] [ComfyUI] Currently allocated : 4.92 GiB [07:54] [ComfyUI] Requested : 286.44 MiB [07:54] [ComfyUI] Device limit : 7.52 GiB [07:54] [ComfyUI] Free (according to CUDA): 11.19 MiB [07:54] [ComfyUI] PyTorch limit (set by user-supplied memory fraction) [07:54] [ComfyUI] : 17179869184.00 GiB [07:54] [ComfyUI] [07:54] [ComfyUI] [07:54] [ComfyUI] Prompt executed in 41.64 seconds [07:54] Execution complete (t=42.2s) [07:54] Execution error: {'prompt_id': '0f8f770a-2093-4692-88a4-c73e63dc11b5', 'node_id': '6', 'node_type': 'DepthAnythingV3_Streaming', 'executed': ['3', '5', '2'], 'exception_message': 'Allocation on device 0 would exceed allowed memory. (out of memory)\nCurrently allocated : 4.92 GiB\nRequested : 286.44 MiB\nDevice limit : 7.52 GiB\nFree (according to CUDA): 11.19 MiB\nPyTorch limit (set by user-supplied memory fraction)\n : 17179869184.00 GiB\n\nWorker traceback:\nTraceback (most recent call last):\n File "/tmp/comfyui_pvenv_4rmut23q/persistent_worker.py", line 1762, in main\n result = method(**inputs)\n ^^^^^^^^^^^^^^^^\n File "/workspaces/DepthAnythingV3-0307/ComfyUI/comfy_api/latest/_io.py", line 1833, in EXECUTE_NORMALIZED\n to_return = cls.execute(*args, **kwargs)\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/streaming/node.py", line 249, in execute\n result = pipeline.run(normalized_images, pbar=pbar,\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/streaming/pipeline.py", line 277, in run\n result = self._process_chunk(normalized_images, start, end)\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/streaming/pipeline.py", line 330, in _process_chunk\n output = self.model(chunk_input)\n ^^^^^^^^^^^^^^^^^^^^^^^\n File "/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1779, in _wrapped_call_impl\n return self._call_impl(*args, **kwargs)\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n File "/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1790, in _call_impl\n return forward_call(*args, **kwargs)\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/depth_anything_v3/model.py", line 1577, in forward\n output = self._process_depth_head(feats, H, W)\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/depth_anything_v3/model.py", line 1591, in _process_depth_head\n return self.head(feats, H, W, patch_start_idx=0)\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n File "/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1779, in _wrapped_call_impl\n return self._call_impl(*args, **kwargs)\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n File "/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1790, in _call_impl\n return forward_call(*args, **kwargs)\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/depth_anything_v3/model.py", line 1379, in forward\n self._forward_impl([feat[s0:s1] for feat in feats], H, W, patch_start_idx)\n File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/depth_anything_v3/model.py", line 1419, in _forward_impl\n fused_main = self._add_pos_embed(fused_main, W, H)\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/depth_anything_v3/model.py", line 1477, in _add_pos_embed\n pe = position_grid_to_embed(pe, x.shape[1]) * ratio\n ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~\ntorch.OutOfMemoryError: Allocation on device 0 would exceed allowed memory. (out of memory)\nCurrently allocated : 4.92 GiB\nRequested : 286.44 MiB\nDevice limit : 7.52 GiB\nFree (according to CUDA): 11.19 MiB\nPyTorch limit (set by user-supplied memory fraction)\n : 17179869184.00 GiB\n\n', 'exception_type': 'comfy_env.isolation.workers.base.WorkerError', 'traceback': [' File "/workspaces/DepthAnythingV3-0307/ComfyUI/execution.py", line 535, in execute\n output_data, output_ui, has_subgraph, has_pending_tasks = await get_output_data(prompt_id, unique_id, obj, input_data_all, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, v3_data=v3_data)\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n', ' File "/workspaces/DepthAnythingV3-0307/ComfyUI/execution.py", line 335, in get_output_data\n return_values = await _async_map_node_over_list(prompt_id, unique_id, obj, input_data_all, obj.FUNCTION, allow_interrupt=True, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, v3_data=v3_data)\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n', ' File "/workspaces/DepthAnythingV3-0307/ComfyUI/execution.py", line 309, in _async_map_node_over_list\n await process_inputs(input_dict, i)\n', ' File "/workspaces/DepthAnythingV3-0307/ComfyUI/execution.py", line 297, in process_inputs\n result = f(**inputs)\n ^^^^^^^^^^^\n', ' File "/workspaces/DepthAnythingV3-0307/.venv/lib/python3.11/site-packages/comfy_env/isolation/metadata.py", line 536, in proxy\n result = worker.call_method(\n ^^^^^^^^^^^^^^^^^^^\n', ' File "/workspaces/DepthAnythingV3-0307/.venv/lib/python3.11/site-packages/comfy_env/isolation/workers/subprocess.py", line 674, in call_method\n raise WorkerError(\n'], 'current_inputs': {'normalization_mode': ['Standard'], 'chunk_size': [30], 'overlap': [8], 'align_lib': ['auto'], 'align_method': ['sim3'], 'resize_method': ['resize'], 'invert_depth': [True], 'save_pointcloud': [False], 'sample_ratio': [0.015], 'conf_threshold_coef': [0.75], 'da3_model': ["{'model_path': '/workspaces/DepthAnythingV3-0307/ComfyUI/models/depthanything3/da3_large.safetensors', 'model_key': 'da3-large', 'dtype': 'bf16', 'attention': 'auto'}"], 'video': [''], 'salad_model': ["{'checkpoint_path': '/workspaces/DepthAnythingV3-0307/ComfyUI/models/salad/dino_salad.safetensors'}"]}, 'current_outputs': ['4', '6', '3', '5', '2'], 'timestamp': 1778382949205} [07:55] Capturing execution frames: simple.json [08:00] Frame 1 saved (frame_000.jpg, t=0.0s) [08:00] Validating workflow... [08:00] [ComfyUI] === /validate endpoint called === [08:00] [ComfyUI] Received JSON with keys: ['prompt'] [08:00] [ComfyUI] Prompt has 4 nodes [08:00] [ComfyUI] Node 2: LoadImage [08:00] [ComfyUI] Inputs: {"image": "bridge.jpg"} [08:00] Queuing workflow for execution... [08:00] [ComfyUI] Node 3: DownloadAndLoadDepthAnythingV3Model [08:00] [ComfyUI] Inputs: {"model": "da3_large.safetensors", "precision": "auto", "attention": "auto"} [08:00] [ComfyUI] Node 4: PreviewImage [08:00] [ComfyUI] Inputs: {"images": ["14", 0]} [08:00] [ComfyUI] Node 14: DepthAnything_V3 [08:00] [ComfyUI] Inputs: {"normalization_mode": "V2-Style", "resize_method": "resize", "invert_depth": true, "keep_model_size": false, "da3_model": ["3", 0], "images": ["2", 0]} [08:00] [ComfyUI] Generated prompt_id: 893e3902-2173-41aa-83ef-36097ffe86dc [08:00] [ComfyUI] Calling execution.validate_prompt()... [08:00] [ComfyUI] Validation result: valid=True [08:00] [ComfyUI] === Validation PASSED === [08:00] [ComfyUI] got prompt [08:00] [ComfyUI] [comfy-env] Running DownloadAndLoadDepthAnythingV3Model: model: 'da3_large.safetensors', precision: 'auto', attention: 'auto' [08:00] [ComfyUI] [VRAM] Before DownloadAndLoadDepthAnythingV3Model: 7692 / 7703 MB [08:00] [ComfyUI] [SubprocessWorker] call_method: nodes.load_model.DownloadAndLoadDepthAnythingV3Model.EXECUTE_NORMALIZED [08:00] [ComfyUI] [SubprocessWorker] acquired lock, ensuring started... [08:00] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [08:00] [ComfyUI] [ComfyUI-DepthAnythingV3] python: /workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/bin/python [08:00] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: failed [08:00] [ComfyUI] [ComfyUI-DepthAnythingV3] Socket unhealthy, restarting worker... [08:00] [ComfyUI] [ComfyUI-DepthAnythingV3] sys_path sent to worker: ['/workspaces/DepthAnythingV3-0307/ComfyUI', '/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3', '/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/lib/python3.11/site-packages', '/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3'] [08:00] [ComfyUI] [SubprocessWorker] is_pixi=True, python=/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/bin/python [08:00] [ComfyUI] [SubprocessWorker] launching cmd=['/root/.pixi/bin/pixi', 'run', '--as-is', '--manifest-path', '/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/pixi.toml', '-e', 'depthanythingv3-nodes', 'python', '/tmp/comfyui_pvenv_4rmut23q/persistent_worker.py', 'abstract://comfy_worker_c352708d8e9c']... [08:00] [ComfyUI] [SubprocessWorker] socket_addr=abstract://comfy_worker_c352708d8e9c (abstract, no filesystem path) [08:02] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] WARNING: WARNING: You need pytorch with cu130 or higher to use optimized CUDA operations. [08:02] [ComfyUI] [SubprocessWorker] worker started/confirmed [08:02] [ComfyUI] [SubprocessWorker] serializing kwargs to shm... [08:02] [ComfyUI] [SubprocessWorker] created 0 shm blocks [08:02] [ComfyUI] [SubprocessWorker] sending request via socket... [08:02] [ComfyUI] [SubprocessWorker] call_id=30 sending call_method [08:02] Frame 2 saved (frame_001.jpg, t=2.3s) [08:02] [capture-periodic] freeze=1ms shot=145ms unfreeze=4ms saved=True [08:03] [ComfyUI] [SubprocessWorker] got response: ok [08:03] [ComfyUI] [comfy-env] DownloadAndLoadDepthAnythingV3Model done (2.40s): result: NodeOutput [08:03] [ComfyUI] [VRAM] After DownloadAndLoadDepthAnythingV3Model: 386 / 7703 MB [08:03] [ComfyUI] [comfy-env] Running DepthAnything_V3: normalization_mode: 'V2-Style', resize_method: 'resize', invert_depth: True, keep_model_size: False, da3_model: dict, images: torch.float32 [1,683,1024,3] cpu [08:03] [ComfyUI] [VRAM] Before DepthAnything_V3: 386 / 7703 MB [08:03] [ComfyUI] [SubprocessWorker] call_method: nodes.nodes_inference.DepthAnything_V3.EXECUTE_NORMALIZED [08:03] [ComfyUI] [SubprocessWorker] acquired lock, ensuring started... [08:03] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [08:03] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ok [08:03] [ComfyUI] [SubprocessWorker] worker started/confirmed [08:03] [ComfyUI] [comfy-env] PRE-SERIALIZE 'images' shape: torch.Size([1, 683, 1024, 3]) [08:03] [ComfyUI] [SubprocessWorker] serializing kwargs to shm... [08:03] [ComfyUI] [SubprocessWorker] created 0 shm blocks [08:03] [ComfyUI] [SubprocessWorker] sending request via socket... [08:03] [ComfyUI] [SubprocessWorker] call_id=31 sending call_method [08:03] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [08:03] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[1, 683, 1024, 3] [08:03] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([1, 683, 1024, 3]) [08:04] [ComfyUI] [SubprocessWorker] callback 'request_vram_budget' from call_id=31 [08:04] [ComfyUI] [comfy-env] VRAM request: 0.82GB, free before eviction: 7.67GB [08:04] [ComfyUI] [comfy-env] VRAM after eviction: 7.67GB [08:04] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +554MB (now 554MB) reserved=576MB peak=554MB [08:04] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] WARNING: WARNING: This model does not support sky segmentation. V2-Style normalization will work but without sky masking. Use Mono/Metric/Nested models for best V2-Style results. [08:05] Frame 3 saved (frame_002.jpg, t=4.6s) [08:05] [capture-periodic] freeze=2ms shot=152ms unfreeze=5ms saved=True [08:05] [ComfyUI] [SubprocessWorker] callback 'report_progress' from call_id=31 [08:05] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +263MB (now 817MB) reserved=2528MB peak=1917MB [08:05] [ComfyUI] [SubprocessWorker] got response: ok [08:05] [ComfyUI] [comfy-env] Created 1 model patchers: ['DepthAnything3Net_1'] [08:05] [ComfyUI] [comfy-env] DepthAnything_V3 done (2.41s): result: NodeOutput [08:05] [ComfyUI] [VRAM] After DepthAnything_V3: 3022 / 7703 MB [08:05] [ComfyUI] [VRAM] Loaded: SubprocessModel (783 MB) [08:05] [ComfyUI] Prompt executed in 4.87 seconds [08:05] Node executed (1 total), capturing... [08:05] Frame 4 saved (frame_003.jpg, t=5.3s) [08:05] [capture-node] freeze=2ms shot=149ms unfreeze=1ms saved=True [08:05] Execution complete (t=5.6s) [08:05] Captured 4 unique frames over 5.57s [08:06] [debug] iframes: 0 [08:06] [timing] trigger_3d_previews: 0.0s [08:08] Saved high-quality screenshot: simple_executed.png [08:08] [timing] final_screenshot: 0.2s [08:08] [ComfyUI] [comfy-env] [VRAM] Before offload 'DepthAnything3Net_1' (783 MB) -> cpu: 3022 / 7703 MB [08:08] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [08:08] Capturing execution frames: video_multiview_depth.json [08:08] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ok [08:08] [ComfyUI] [SubprocessWorker] call_id=32 sending model_to_device [08:08] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] -202MB (now 615MB) reserved=2528MB peak=1917MB [08:08] [ComfyUI] [comfy-env] [VRAM] After offload 'DepthAnything3Net_1' (783 MB): 558 / 7703 MB [08:13] Frame 1 saved (frame_000.jpg, t=0.0s) [08:13] Validating workflow... [08:13] [ComfyUI] === /validate endpoint called === [08:13] [ComfyUI] Received JSON with keys: ['prompt'] [08:13] [ComfyUI] Prompt has 9 nodes [08:13] [ComfyUI] Node 3: DownloadAndLoadDepthAnythingV3Model [08:13] [ComfyUI] Inputs: {"model": "da3_base.safetensors", "precision": "auto", "attention": "auto"} [08:13] [ComfyUI] Node 52: CreateVideo [08:13] Queuing workflow for execution... [08:13] [ComfyUI] Inputs: {"fps": 10, "images": ["62", 0]} [08:13] [ComfyUI] Node 53: CreateVideo [08:13] [ComfyUI] Inputs: {"fps": 10, "images": ["61", 0]} [08:13] [ComfyUI] Node 54: SaveVideo [08:13] [ComfyUI] Inputs: {"filename_prefix": "video/ComfyUI", "format": "auto", "codec": "auto", "video": ["52", 0]} [08:13] [ComfyUI] Node 55: SaveVideo [08:13] [ComfyUI] Inputs: {"filename_prefix": "video/ComfyUI", "format": "auto", "codec": "auto", "video": ["53", 0]} [08:13] [ComfyUI] Node 61: DepthAnything_V3 [08:13] [ComfyUI] Inputs: {"normalization_mode": "V2-Style", "resize_method": "resize", "invert_depth": false, "keep_model_size": false, "da3_model": ["3", 0], "images": ["71", 0]} [08:13] [ComfyUI] Node 62: DepthAnythingV3_MultiView [08:13] [ComfyUI] Inputs: {"normalization_mode": "V2-Style", "resize_method": "resize", "invert_depth": false, "keep_model_size": false, "da3_model": ["3", 0], "images": ["71", 0]} [08:13] [ComfyUI] Node 66: PreviewImage [08:13] [ComfyUI] Inputs: {"images": ["71", 0]} [08:13] [ComfyUI] Node 71: LoadImagesFromFolderKJ [08:13] [ComfyUI] Inputs: {"folder": "input/robotframes", "width": 1024, "height": 576, "keep_aspect_ratio": "crop", "image_load_cap": 0, "start_index": 0, "include_subfolders": false} [08:13] [ComfyUI] Generated prompt_id: e97c9832-8462-430f-aaf4-487bd0ddcb87 [08:13] [ComfyUI] Calling execution.validate_prompt()... [08:13] [ComfyUI] Validation result: valid=True [08:13] [ComfyUI] === Validation PASSED === [08:13] [ComfyUI] got prompt [08:14] [ComfyUI] [comfy-env] Running DownloadAndLoadDepthAnythingV3Model: model: 'da3_base.safetensors', precision: 'auto', attention: 'auto' [08:14] [ComfyUI] [ComfyUI-DepthAnythingV3] python: /workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/bin/python [08:14] [ComfyUI] [VRAM] Before DownloadAndLoadDepthAnythingV3Model: 558 / 7703 MB [08:14] [ComfyUI] [ComfyUI-DepthAnythingV3] sys_path sent to worker: ['/workspaces/DepthAnythingV3-0307/ComfyUI', '/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3', '/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/lib/python3.11/site-packages', '/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3'] [08:14] [ComfyUI] [SubprocessWorker] call_method: nodes.load_model.DownloadAndLoadDepthAnythingV3Model.EXECUTE_NORMALIZED [08:14] [ComfyUI] [SubprocessWorker] is_pixi=True, python=/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/bin/python [08:14] [ComfyUI] [SubprocessWorker] acquired lock, ensuring started... [08:14] [ComfyUI] [SubprocessWorker] launching cmd=['/root/.pixi/bin/pixi', 'run', '--as-is', '--manifest-path', '/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/pixi.toml', '-e', 'depthanythingv3-nodes', 'python', '/tmp/comfyui_pvenv_4rmut23q/persistent_worker.py', 'abstract://comfy_worker_5b2d72bb4e43']... [08:14] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [08:14] [ComfyUI] [SubprocessWorker] socket_addr=abstract://comfy_worker_5b2d72bb4e43 (abstract, no filesystem path) [08:14] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: failed [08:14] [ComfyUI] [ComfyUI-DepthAnythingV3] Socket unhealthy, restarting worker... [08:14] [ComfyUI] [comfy-env] Invalidated 1 stale model patchers (will be cleaned up during next unload) [08:14] Node executed (1 total), capturing... [08:14] Frame 2 saved (frame_001.jpg, t=1.3s) [08:14] [capture-node] freeze=2ms shot=140ms unfreeze=6ms saved=True [08:16] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] WARNING: WARNING: You need pytorch with cu130 or higher to use optimized CUDA operations. [08:16] [ComfyUI] [SubprocessWorker] worker started/confirmed [08:16] [ComfyUI] [SubprocessWorker] serializing kwargs to shm... [08:16] [ComfyUI] [SubprocessWorker] created 0 shm blocks [08:16] [ComfyUI] [SubprocessWorker] sending request via socket... [08:16] [ComfyUI] [SubprocessWorker] call_id=33 sending call_method [08:17] [ComfyUI] [08:19] Frame 3 saved (frame_002.jpg, t=5.8s) [08:19] [capture-periodic] freeze=2ms shot=160ms unfreeze=20ms saved=True [08:19] [capture-loop] iter=50 t=6.1s state={'complete': False, 'error': None, 'executedCount': 1, 'wsState': 1} eval_ms=2 ram=6.3/31.3GB (20%) swap=0.0/32.0GB (0%) cpu=14% browser_rss=1.22GB [08:21] [ComfyUI] model.safetensors: 0%| | 0.00/542M [00:00 cpu: 386 / 7703 MB [08:23] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [08:23] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ok [08:23] [ComfyUI] [SubprocessWorker] call_id=34 sending model_to_device [08:23] [ComfyUI] Model 'DepthAnything3Net_1' gone from worker (restarted?), treating as offloaded [08:23] [ComfyUI] [comfy-env] [VRAM] After offload 'DepthAnything3Net_1' (783 MB): 386 / 7703 MB [08:23] [ComfyUI] [comfy-env] DownloadAndLoadDepthAnythingV3Model done (8.97s): result: NodeOutput [08:23] [ComfyUI] [VRAM] After DownloadAndLoadDepthAnythingV3Model: 386 / 7703 MB [08:23] [ComfyUI] [comfy-env] Running DepthAnything_V3: normalization_mode: 'V2-Style', resize_method: 'resize', invert_depth: False, keep_model_size: False, da3_model: dict, images: torch.float32 [13,576,1024,3] cpu [08:23] [ComfyUI] [VRAM] Before DepthAnything_V3: 386 / 7703 MB [08:23] [ComfyUI] [SubprocessWorker] call_method: nodes.nodes_inference.DepthAnything_V3.EXECUTE_NORMALIZED [08:23] [ComfyUI] [SubprocessWorker] acquired lock, ensuring started... [08:23] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [08:23] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ok [08:23] [ComfyUI] [SubprocessWorker] worker started/confirmed [08:23] [ComfyUI] [comfy-env] PRE-SERIALIZE 'images' shape: torch.Size([13, 576, 1024, 3]) [08:23] [ComfyUI] [SubprocessWorker] serializing kwargs to shm... [08:23] [ComfyUI] [SubprocessWorker] created 0 shm blocks [08:23] [ComfyUI] [SubprocessWorker] sending request via socket... [08:23] [ComfyUI] [SubprocessWorker] call_id=35 sending call_method [08:23] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [08:23] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[13, 576, 1024, 3] [08:23] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([13, 576, 1024, 3]) [08:24] Frame 5 saved (frame_004.jpg, t=10.4s) [08:24] [capture-periodic] freeze=2ms shot=136ms unfreeze=5ms saved=True [08:24] [ComfyUI] [SubprocessWorker] callback 'request_vram_budget' from call_id=35 [08:24] [ComfyUI] [comfy-env] VRAM request: 0.27GB, free before eviction: 7.67GB [08:24] [ComfyUI] [comfy-env] VRAM after eviction: 7.67GB [08:24] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] WARNING: WARNING: This model does not support sky segmentation. V2-Style normalization will work but without sky masking. Use Mono/Metric/Nested models for best V2-Style results. [08:24] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +258MB (now 258MB) reserved=288MB peak=258MB [08:25] [capture-loop] iter=100 t=11.7s state={'complete': False, 'error': None, 'executedCount': 1, 'wsState': 1} eval_ms=2 ram=7.5/31.3GB (24%) swap=0.0/32.0GB (0%) cpu=16% browser_rss=1.23GB [08:25] [ComfyUI] [SubprocessWorker] callback 'report_progress' from call_id=35 [08:25] [ComfyUI] [SubprocessWorker] callback 'report_progress' from call_id=35 [08:25] [ComfyUI] [SubprocessWorker] callback 'report_progress' from call_id=35 [08:25] [ComfyUI] [SubprocessWorker] callback 'report_progress' from call_id=35 [08:25] [ComfyUI] [SubprocessWorker] callback 'report_progress' from call_id=35 [08:25] [ComfyUI] [SubprocessWorker] callback 'report_progress' from call_id=35 [08:25] [ComfyUI] [SubprocessWorker] callback 'report_progress' from call_id=35 [08:26] [ComfyUI] [SubprocessWorker] callback 'report_progress' from call_id=35 [08:26] [ComfyUI] [SubprocessWorker] callback 'report_progress' from call_id=35 [08:26] Frame 6 saved (frame_005.jpg, t=12.6s) [08:26] [capture-periodic] freeze=6ms shot=145ms unfreeze=6ms saved=True [08:26] [ComfyUI] [SubprocessWorker] callback 'report_progress' from call_id=35 [08:26] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +378MB (now 637MB) reserved=1120MB peak=637MB [08:26] [ComfyUI] [SubprocessWorker] callback 'report_progress' from call_id=35 [08:26] [ComfyUI] [SubprocessWorker] callback 'report_progress' from call_id=35 [08:26] [ComfyUI] [SubprocessWorker] callback 'report_progress' from call_id=35 [08:27] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] -350MB (now 287MB) reserved=1120MB peak=820MB [08:28] [ComfyUI] [SubprocessWorker] got response: ok [08:28] [ComfyUI] [comfy-env] Created 1 model patchers: ['DepthAnything3Net_1'] [08:28] [ComfyUI] [comfy-env] DepthAnything_V3 done (4.67s): result: NodeOutput [08:28] [ComfyUI] [VRAM] After DepthAnything_V3: 1614 / 7703 MB [08:28] [ComfyUI] [VRAM] Loaded: SubprocessModel (258 MB) [08:28] [ComfyUI] [comfy-env] Running DepthAnythingV3_MultiView: normalization_mode: 'V2-Style', resize_method: 'resize', invert_depth: False, keep_model_size: False, da3_model: dict, images: torch.float32 [13,576,1024,3] cpu [08:28] [ComfyUI] [VRAM] Before DepthAnythingV3_MultiView: 1614 / 7703 MB [08:28] [ComfyUI] [VRAM] Loaded: SubprocessModel (258 MB) [08:28] [ComfyUI] [SubprocessWorker] call_method: nodes.nodes_multiview.DepthAnythingV3_MultiView.EXECUTE_NORMALIZED [08:28] [ComfyUI] [SubprocessWorker] acquired lock, ensuring started... [08:28] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [08:28] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ok [08:28] [ComfyUI] [SubprocessWorker] worker started/confirmed [08:28] [ComfyUI] [comfy-env] PRE-SERIALIZE 'images' shape: torch.Size([13, 576, 1024, 3]) [08:28] [ComfyUI] [SubprocessWorker] serializing kwargs to shm... [08:28] [ComfyUI] [SubprocessWorker] created 0 shm blocks [08:28] [ComfyUI] [SubprocessWorker] sending request via socket... [08:28] [ComfyUI] [SubprocessWorker] call_id=36 sending call_method [08:28] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] _from_shm got dict with keys: ['__type__', 'strategy', 'parent_pid', 'fd', 'storage_size'] [08:28] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZE TensorRef: tensor_size=[13, 576, 1024, 3] [08:28] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [comfy-env] DESERIALIZED tensor shape: torch.Size([13, 576, 1024, 3]) [08:28] [ComfyUI] [SubprocessWorker] callback 'request_vram_budget' from call_id=36 [08:28] [ComfyUI] [comfy-env] VRAM request: 0.27GB, free before eviction: 6.38GB [08:28] [ComfyUI] [comfy-env] VRAM after eviction: 6.38GB [08:28] Frame 7 saved (frame_006.jpg, t=14.8s) [08:28] [capture-periodic] freeze=1ms shot=107ms unfreeze=6ms saved=True [08:28] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] WARNING: WARNING: This model does not support sky segmentation. V2-Style normalization will work but without sky masking. Use Mono/Metric/Nested models for best V2-Style results. [08:28] Node executed (2 total), capturing... [08:28] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +423MB (now 711MB) reserved=1088MB peak=820MB [08:28] Frame 8 saved (frame_007.jpg, t=15.2s) [08:28] [capture-node] freeze=10ms shot=150ms unfreeze=7ms saved=True [08:29] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] +726MB (now 1437MB) reserved=1984MB peak=1437MB [08:31] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] [VRAM] -784MB (now 653MB) reserved=2848MB peak=1458MB [08:31] Frame 9 saved (frame_008.jpg, t=17.5s) [08:31] [capture-periodic] freeze=1ms shot=137ms unfreeze=5ms saved=True [08:31] [capture-loop] iter=150 t=17.8s state={'complete': False, 'error': None, 'executedCount': 2, 'wsState': 1} eval_ms=2 ram=8.8/31.3GB (28%) swap=0.0/32.0GB (0%) cpu=18% browser_rss=1.08GB [08:31] [ComfyUI] [SubprocessWorker] got response: ok [08:31] [ComfyUI] [comfy-env] DepthAnythingV3_MultiView done (3.53s): result: NodeOutput [08:31] [ComfyUI] [VRAM] After DepthAnythingV3_MultiView: 3342 / 7703 MB [08:31] [ComfyUI] [VRAM] Loaded: SubprocessModel (258 MB) [08:32] [ComfyUI] Prompt executed in 18.45 seconds [08:32] [ComfyUI] [comfy-env] [VRAM] Before offload 'DepthAnything3Net_1' (258 MB) -> cpu: 3342 / 7703 MB [08:32] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: ping (timeout=5.0s)... [08:32] [ComfyUI] [ComfyUI-DepthAnythingV3] Health check: failed [08:32] [ComfyUI] [ComfyUI-DepthAnythingV3] Socket unhealthy, restarting worker... [08:32] [ComfyUI] [comfy-env] Invalidated 1 stale model patchers (will be cleaned up during next unload) [08:32] [ComfyUI] [ComfyUI-DepthAnythingV3] python: /workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/bin/python [08:32] [ComfyUI] [ComfyUI-DepthAnythingV3] sys_path sent to worker: ['/workspaces/DepthAnythingV3-0307/ComfyUI', '/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3', '/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/lib/python3.11/site-packages', '/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3'] [08:32] [ComfyUI] [SubprocessWorker] is_pixi=True, python=/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/bin/python [08:32] [ComfyUI] [SubprocessWorker] launching cmd=['/root/.pixi/bin/pixi', 'run', '--as-is', '--manifest-path', '/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/pixi.toml', '-e', 'depthanythingv3-nodes', 'python', '/tmp/comfyui_pvenv_4rmut23q/persistent_worker.py', 'abstract://comfy_worker_02f05d624d15']... [08:32] [ComfyUI] [SubprocessWorker] socket_addr=abstract://comfy_worker_02f05d624d15 (abstract, no filesystem path) [08:32] Node executed (3 total), capturing... [08:32] Frame 10 saved (frame_009.jpg, t=18.9s) [08:32] [capture-node] freeze=2ms shot=141ms unfreeze=0ms saved=True [08:32] Execution complete (t=18.9s) [08:32] Captured 10 unique frames over 19.06s [08:32] [debug] iframes: 0 [08:32] [timing] trigger_3d_previews: 0.0s [08:34] [ComfyUI] [worker:ComfyUI-DepthAnythingV3] WARNING: WARNING: You need pytorch with cu130 or higher to use optimized CUDA operations. [08:34] [ComfyUI] [SubprocessWorker] call_id=37 sending model_to_device [08:34] [ComfyUI] Model 'DepthAnything3Net_1' gone from worker (restarted?), treating as offloaded [08:34] [ComfyUI] [comfy-env] [VRAM] After offload 'DepthAnything3Net_1' (258 MB): 386 / 7703 MB [08:34] Saved high-quality screenshot: video_multiview_depth_executed.png [08:34] [timing] final_screenshot: 0.3s [08:35] Results saved to /logs/DepthAnythingV3-0307/dev/linux-gpu/results.json [08:35] Model report: 9 files, 2.4 GB -> /logs/DepthAnythingV3-0307/dev/linux-gpu/models.json [08:35] Saved: /logs/DepthAnythingV3-0307/dev/linux-gpu/index.html [08:35] linux: FAILED [08:35] Error: Workflow execution failed (1 error(s)) [08:35] Details: da3_streaming.json: Workflow execution failed: {'prompt_id': '0f8f770a-2093-4692-88a4-c73e63dc11b5', 'node_id': '6', 'node_type': 'DepthAnythingV3_Streaming', 'executed': ['3', '5', '2'], 'exception_message': 'Allocation on device 0 would exceed allowed memory. (out of memory)\nCurrently allocated : 4.92 GiB\nRequested : 286.44 MiB\nDevice limit : 7.52 GiB\nFree (according to CUDA): 11.19 MiB\nPyTorch limit (set by user-supplied memory fraction)\n : 17179869184.00 GiB\n\nWorker traceback:\nTraceback (most recent call last):\n File "/tmp/comfyui_pvenv_4rmut23q/persistent_worker.py", line 1762, in main\n result = method(**inputs)\n ^^^^^^^^^^^^^^^^\n File "/workspaces/DepthAnythingV3-0307/ComfyUI/comfy_api/latest/_io.py", line 1833, in EXECUTE_NORMALIZED\n to_return = cls.execute(*args, **kwargs)\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/streaming/node.py", line 249, in execute\n result = pipeline.run(normalized_images, pbar=pbar,\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/streaming/pipeline.py", line 277, in run\n result = self._process_chunk(normalized_images, start, end)\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/streaming/pipeline.py", line 330, in _process_chunk\n output = self.model(chunk_input)\n ^^^^^^^^^^^^^^^^^^^^^^^\n File "/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1779, in _wrapped_call_impl\n return self._call_impl(*args, **kwargs)\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n File "/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1790, in _call_impl\n return forward_call(*args, **kwargs)\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/depth_anything_v3/model.py", line 1577, in forward\n output = self._process_depth_head(feats, H, W)\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/depth_anything_v3/model.py", line 1591, in _process_depth_head\n return self.head(feats, H, W, patch_start_idx=0)\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n File "/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1779, in _wrapped_call_impl\n return self._call_impl(*args, **kwargs)\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n File "/workspaces/DepthAnythingV3-0307/ComfyUI/.ce/.pixi/envs/depthanythingv3-nodes/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1790, in _call_impl\n return forward_call(*args, **kwargs)\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/depth_anything_v3/model.py", line 1379, in forward\n self._forward_impl([feat[s0:s1] for feat in feats], H, W, patch_start_idx)\n File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/depth_anything_v3/model.py", line 1419, in _forward_impl\n fused_main = self._add_pos_embed(fused_main, W, H)\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n File "/workspaces/DepthAnythingV3-0307/ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/nodes/depth_anything_v3/model.py", line 1477, in _add_pos_embed\n pe = position_grid_to_embed(pe, x.shape[1]) * ratio\n ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~\ntorch.OutOfMemoryError: Allocation on device 0 would exceed allowed memory. (out of memory)\nCurrently allocated : 4.92 GiB\nRequested : 286.44 MiB\nDevice limit : 7.52 GiB\nFree (according to CUDA): 11.19 MiB\nPyTorch limit (set by user-supplied memory fraction)\n : 17179869184.00 GiB\n\n', 'exception_type': 'comfy_env.isolation.workers.base.WorkerError', 'traceback': [' File "/workspaces/DepthAnythingV3-0307/ComfyUI/execution.py", line 535, in execute\n output_data, output_ui, has_subgraph, has_pending_tasks = await get_output_data(prompt_id, unique_id, obj, input_data_all, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, v3_data=v3_data)\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n', ' File "/workspaces/DepthAnythingV3-0307/ComfyUI/execution.py", line 335, in get_output_data\n return_values = await _async_map_node_over_list(prompt_id, unique_id, obj, input_data_all, obj.FUNCTION, allow_interrupt=True, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, v3_data=v3_data)\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n', ' File "/workspaces/DepthAnythingV3-0307/ComfyUI/execution.py", line 309, in _async_map_node_over_list\n await process_inputs(input_dict, i)\n', ' File "/workspaces/DepthAnythingV3-0307/ComfyUI/execution.py", line 297, in process_inputs\n result = f(**inputs)\n ^^^^^^^^^^^\n', ' File "/workspaces/DepthAnythingV3-0307/.venv/lib/python3.11/site-packages/comfy_env/isolation/metadata.py", line 536, in proxy\n result = worker.call_method(\n ^^^^^^^^^^^^^^^^^^^\n', ' File "/workspaces/DepthAnythingV3-0307/.venv/lib/python3.11/site-packages/comfy_env/isolation/workers/subprocess.py", line 674, in call_method\n raise WorkerError(\n'], 'current_inputs': {'normalization_mode': ['Standard'], 'chunk_size': [30], 'overlap': [8], 'align_lib': ['auto'], 'align_method': ['sim3'], 'resize_method': ['resize'], 'invert_depth': [True], 'save_pointcloud': [False], 'sample_ratio': [0.015], 'conf_threshold_coef': [0.75], 'da3_model': ["{'model_path': '/workspaces/DepthAnythingV3-0307/ComfyUI/models/depthanything3/da3_large.safetensors', 'model_key': 'da3-large', 'dtype': 'bf16', 'attention': 'auto'}"], 'video': [''], 'salad_model': ["{'checkpoint_path': '/workspaces/DepthAnythingV3-0307/ComfyUI/models/salad/dino_salad.safetensors'}"]}, 'current_outputs': ['4', '6', '3', '5', '2'], 'timestamp': 1778382949205} Details: Workflow: /node/ComfyUI-DepthAnythingV3/workflows/da3_streaming.json Node error: DepthAnythingV3_Streaming [08:35] Stopping ComfyUI server...