Episode 9 , ffmpeg return error

[maxya]

using container gyoridavid/narrated-story-creator:latest-cuda with workflow from episode 9

narrator-story | Using CUDA
narrator-story | INFO: Started server process [1]
narrator-story | INFO: Waiting for application startup.
narrator-story | INFO: Application startup complete.
narrator-story | INFO: Uvicorn running on http://0.0.0.0:8000 (Press CTRL+C to quit)
narrator-story | Saved 1 videos to persistent storage
narrator-story | INFO: 192.168.11.5:44894 - "POST /api/videos HTTP/1.1" 200 OK
narrator-story | Saved 1 videos to persistent storage
narrator-story | Downloading background video for 8951b0fe-7283-4259-ac27-b17620f63a96
narrator-story | Downloading person image for 8951b0fe-7283-4259-ac27-b17620f63a96
narrator-story | creating overlay
narrator-story | Text dimensions: 373x52, Y offset: 24
narrator-story | Overlay saved to /app/tmp/8951b0fe-7283-4259-ac27-b17620f63a96/overlay.png
narrator-story | creating narration
narrator-story | WARNING: Defaulting repo_id to hexgrad/Kokoro-82M. Pass repo_id='hexgrad/Kokoro-82M' to suppress this warning.
narrator-story | /usr/local/lib/python3.10/dist-packages/torch/nn/modules/rnn.py:123: UserWarning: dropout option adds dropout after all but last recurrent layer, so non-zero dropout expects num_layers greater than 1, but got dropout=0.2 and num_layers=1
narrator-story | warnings.warn(
narrator-story | /usr/local/lib/python3.10/dist-packages/torch/nn/utils/weight_norm.py:143: FutureWarning: torch.nn.utils.weight_norm is deprecated in favor of torch.nn.utils.parametrizations.weight_norm.
narrator-story | WeightNorm.apply(module, name, dim)
narrator-story | Collecting en-core-web-sm==3.8.0
narrator-story | Downloading https://github.com/explosion/spacy-models/releases/download/en_core_web_sm-3.8.0/en_core_web_sm-3.8.0-py3-none-any.whl (12.8 MB)
narrator-story | ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 12.8/12.8 MB 17.4 MB/s eta 0:00:00
narrator-story | Installing collected packages: en-core-web-sm
narrator-story | Successfully installed en-core-web-sm-3.8.0
narrator-story | WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv
narrator-story | ✔ Download and installation successful
narrator-story | You can now load the package via spacy.load('en_core_web_sm')
narrator-story | creating subtitle
narrator-story | Subtitle file created: /app/tmp/8951b0fe-7283-4259-ac27-b17620f63a96/subtitle.srt
narrator-story | rendering video
narrator-story | ffmpeg: Error while filtering: Operation not permitted
narrator-story | ffmpeg: frame= 0 fps=0.0 q=0.0 Lsize= 0kB time=N/A bitrate=N/A speed=N/A
narrator-story | ffmpeg: Conversion failed!

[gyoridavid]

Hey @maxya , did you look at the content of the subtitle and the audio? I have a feeling based on your logs, that it was empty. If you ran the workflow, make sure to check the workflow logs, what data was sent to the server.

[maxya]

Hi @gyoridavid ,

I actually did. All data was populated correctly.

I've also tested by shutdown GPU container and started CPU version, node ran without any issues on CPU, then I switched back to the GPU container and re-ran the node with the same context, it failed with the same error.

It feels like ffmpeg does not like the context.
Quick Slack-overflow-ing shows that some of the ffmpeg flags might be an issue.

Any way I can see the Dockerfile?

I can't find repo for this image in your repositories.

Thanks

[gyoridavid]

Hmm..

Could you please add some more context (post data from the nodes) so I can replicate the behavior, that would be helpful.

I didn't open source this tool as I (mostly) merged it into the https://github.com/gyoridavid/ai-agents-no-code-tools

[maxya]

Not super sure about how to include all post data the best way, here's a link to exported workflow with pinned nodes (with data) if this helps.
Download workflow

Since you merged most of this to this repo , I'll probably going to be migrating workflow to use it then.