Run pip install -e .. in path=model/model_training, mac m1 meet error #3352
Replies: 3 comments 1 reply
-
I don't think any of our ML code is intended to run on any device other than Nvidia GPUs. |
Beta Was this translation helpful? Give feedback.
-
IC, thank you |
Beta Was this translation helpful? Give feedback.
-
@OrangeJessie Getting similar errors on google colab. I have set GPU in runtime (Nvidia Tesla T4). I believe as per your earlier comment, it should have worked on above google colab notebook . I am adding the link to my notebook for reference: https://colab.research.google.com/drive/1guc1y2CBzA4BN6h6XHVZYLZlIr1JLSIx?usp=share_link Hoping to get a reply. Thank you. |
Beta Was this translation helpful? Give feedback.
-
Here is the issue, seems like GPU can't find. How can I solve this issue?
I read the historical issues for mac, but don't get anything helpful.
error: subprocess-exited-with-error
× python setup.py egg_info did not run successfully.
│ exit code: 1
╰─> [16 lines of output]
Traceback (most recent call last):
File "", line 2, in
File "", line 34, in
File "/private/var/folders/1y/s1tgwg0d4rqc_nrfwm2f_r140000gp/T/pip-install-u3qof7be/flash-attn_7f22639ed7b743ae9d3555e587d38c77/setup.py", line 83, in
_, bare_metal_major, bare_metal_minor = get_cuda_bare_metal_version(CUDA_HOME)
File "/private/var/folders/1y/s1tgwg0d4rqc_nrfwm2f_r140000gp/T/pip-install-u3qof7be/flash-attn_7f22639ed7b743ae9d3555e587d38c77/setup.py", line 23, in get_cuda_bare_metal_version
raw_output = subprocess.check_output([cuda_dir + "/bin/nvcc", "-V"], universal_newlines=True)
TypeError: unsupported operand type(s) for +: 'NoneType' and 'str'
note: This error originates from a subprocess, and is likely not a problem with pip.
error: metadata-generation-failed
Beta Was this translation helpful? Give feedback.
All reactions