pachyderm / ambient-music-generation Goto Github PK

Hello team and @the-laughing-monkey, @JimmyWhitaker

Not sure if I'm allowed to open an issue for this repo but I thought this question may serve others in the future who use this repo for personal projects. I was able to successfully run the repo itself using the provided trained models and sample music, that worked nicely.

Now I'm trying to generate my own ambient music tunes and I'm working through the steps outlined in this repo under Training Your Own Music Generator in Pachyderm. I downloaded a 3-4min basic sample ambient music tune in hopes of using it to train the models.

I'm stuck on Step 7 "Train the transformer". Following the steps for Step 7, I created a new file called music-transformer.json as required and copied & pasted the json configurations provided here: https://github.com/pachyderm/ambient-music-generation/blob/master/pipelines/music-transformer.json.

I then pushed out the file to PachHub using the pachctl create pipeline -f ./music-transformer.json command. It looks like it was successfully pushed and the repo was created but no job seems to be running for the 'music transformer'. Did I miss any other steps? Any advice appreciated 👍

PS C:\Program Files\Docker\Docker> docker pull jimmywhitaker/music-transformer:0.1
0.1: Pulling from jimmywhitaker/music-transformer
image operating system "linux" cannot be used on this platform

Just wondering if there is a Windows version, thanks!

When following the readme I'm unable to use the provided trained models:

I've cloned the repo, changed to the directory and executed the command:

docker run -it -v pwd\:/data --entrypoint python jimmywhitaker/music-transformer:0.1 generate.py --load_path=/data/trained-models/ambient-musictransformer-model-2-june-3-2020-750-epochs-more-data --inputs /data/samples/midi/1.mid --length=2048 --save_path=/data/generated-music/my-sample-song-1.mid

I get the following FileNotFoundError:

PS D:\ambient-music-generation> docker run -it -v pwd:/data --entrypoint python jimmywhitaker/music-transformer:0.1 generate.py --load_path=/data/trained-models/ambient-musictransformer-model-2-june-3-2020-750-epochs-more-data --inputs /data/samples/midi/1.mid --length=2048 --save_path=/data/generated-music/my-sample-song-1.mid
2022-11-10 21:32:26.830660: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcuda.so.1
2022-11-10 21:32:26.830712: E tensorflow/stream_executor/cuda/cuda_driver.cc:313] failed call to cuInit: UNKNOWN ERROR (-1)
2022-11-10 21:32:26.830736: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:156] kernel driver does not appear to be running on this host (f5005090f1e4): /proc/driver/nvidia/version does not exist
2022-11-10 21:32:26.830944: I tensorflow/core/platform/cpu_feature_guard.cc:143] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA
2022-11-10 21:32:26.838405: I tensorflow/core/platform/profile_utils/cpu_utils.cc:102] CPU Frequency: 2904010000 Hz
2022-11-10 21:32:26.839976: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x7f271c000b20 initialized for platform Host (this does not guarantee that XLA will be used). Devices:
2022-11-10 21:32:26.840017: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): Host, Default Version

generate with decoder wise... beam size is None
Traceback (most recent call last):
File "generate.py", line 52, in
mt = MusicTransformerDecoder(loader_path=load_path)
File "/src/model.py", line 267, in init
self.load_config_file(loader_path)
File "/src/model.py", line 366, in load_config_file
with open(config_path, 'r') as f:
FileNotFoundError: [Errno 2] No such file or directory: '/data/trained-models/ambient-musictransformer-model-2-june-3-2020-750-epochs-more-data/config.json'
PS D:\ambient-music-generation>

Hi,
Great work! I want to try to automate generating ambient sound effects based on the output midi file.
But I had a hard time finding such ambient instruments, can you recommend some of these?