kenwaytis / faster-SadTalker-API

The API server version of the SadTalker project. Runs in Docker, 10 times faster than the original!
MIT License
109 stars 18 forks source link

Errors on image from docker hub #12

Open HasseM opened 6 months ago

HasseM commented 6 months ago

When I just take the image from docker hub, it should work without any additional steps, right?

This is what I get:

`2024-01-03 19:06:44

2024-01-03 19:06:44 ==========

2024-01-03 19:06:44 == CUDA ==

2024-01-03 19:06:44 ==========

2024-01-03 19:06:44 2024-01-03 19:06:44 CUDA Version 11.8.0 2024-01-03 19:06:44 2024-01-03 19:06:44 Container image Copyright (c) 2016-2023, NVIDIA CORPORATION & AFFILIATES. All rights reserved. 2024-01-03 19:06:44 2024-01-03 19:06:44 This container image and its contents are governed by the NVIDIA Deep Learning Container License. 2024-01-03 19:06:44 By pulling and using the container, you accept the terms and conditions of this license: 2024-01-03 19:06:44 https://developer.nvidia.com/ngc/nvidia-deep-learning-container-license 2024-01-03 19:06:44 2024-01-03 19:06:44 A copy of this license is made available in this container at /NGC-DL-CONTAINER-LICENSE for your convenience. 2024-01-03 19:06:44 2024-01-03 19:06:44 WARNING: The NVIDIA Driver was not detected. GPU functionality will not be available. 2024-01-03 19:06:44 Use the NVIDIA Container Toolkit to start this container with GPU support; see 2024-01-03 19:06:44 https://docs.nvidia.com/datacenter/cloud-native/ . 2024-01-03 19:06:44 2024-01-03 19:06:50 font_manager.py :1547 2024-01-03 18:06:50,240 generated new fontManager 2024-01-03 19:06:50 Traceback (most recent call last): 2024-01-03 19:06:50 File "/handler.py", line 8, in 2024-01-03 19:06:50 from inference import main 2024-01-03 19:06:50 File "/inference.py", line 9, in 2024-01-03 19:06:50 from src.facerender.animate import AnimateFromCoeff 2024-01-03 19:06:50 File "/src/facerender/animate.py", line 27, in 2024-01-03 19:06:50 from utils import upload_firebase_file, save_video_without_watermark 2024-01-03 19:06:50 File "/utils.py", line 31, in 2024-01-03 19:06:50 load_credentials()
2024-01-03 19:06:50 File "/utils.py", line 17, in load_credentials 2024-01-03 19:06:50 "project_id": os.environ["project_id"], 2024-01-03 19:06:50 File "/usr/lib/python3.8/os.py", line 675, in getitem 2024-01-03 19:06:50 raise KeyError(key) from None 2024-01-03 19:06:50 KeyError: 'project_id'`

I have an RTX3090 and all NVideo drivers installed but I am on Win11. I'm really looking for a way to get ST working through an API or commandline. I have it running as a SD extension but no way to externally control it that way. So I was looking for running solutions in Docker with an API. Found yours. But these ST docker images seem to have problems with nvidia?

eoffermann commented 3 months ago

Looks like you haven't installed the NVidia Container Toolkit - https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/install-guide.html