Could not solve for environment specs #27
Comments
|
I think the problem is that I kept platform-specific build constraints in the |
|
Yeah I'm on that git hash already. |
|
@kkoutini still having this issue :( |
|
Hi Joseph,
conda create -n ba3l python=3.8
conda activate ba3l
# MPS acceleration is available on MacOS 12.3+
pip3 install torch torchvision torchaudiocheck if torch is installed properly: import torch
import math
# this ensures that the current MacOS version is at least 12.3+
print(torch.backends.mps.is_available())
# this ensures that the current current PyTorch installation was built with MPS activated.
print(torch.backends.mps.is_built())
pip3 install timm==0.4.12At this point inference should work on your machine, you can test it using the (HEAR repo)[https://github.com/kkoutini/passt_hear21]: pip install -e 'git+https://github.com/kkoutini/passt_hear21@0.0.19#egg=hear21passt' then test the inference like this: import torch
from hear21passt.base import load_model, get_scene_embeddings, get_timestamp_embeddings
device = torch.device("mps")
model = load_model().to(device)
seconds = 15
audio = torch.ones((3, 32000 * seconds))*0.5
embed, time_stamps = get_timestamp_embeddings(audio, model)
print(embed.shape)
embed = get_scene_embeddings(audio, model)
print(embed.shape)
# custom dependencies
pip3 install -e 'git+https://github.com/kkoutini/ba3l@v0.0.2#egg=ba3l'
pip3 install -e 'git+https://github.com/kkoutini/pytorch-lightning@v0.0.1#egg=pytorch-lightning'
pip3 install -e 'git+https://github.com/kkoutini/sacred@v0.0.1#egg=sacred' I think now you can run experiments as described in the main repo |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
I clone the repo. As per the README:
But then the next mambo command fails :\
with
This is on an OSX Apple Silicon machine
The text was updated successfully, but these errors were encountered: