python: command not found #20

New Issue

mrchucksneed · 2023-02-21T20:16:51Z

mrchucksneed commented

2023-02-21 20:16:51 +00:00

Hello, me again.
I've got Ubuntu set up now, got ROCM installed and AFAIK working, but I get command not found errors when attempting to run the install script:

> ~/ai-voice-cloning$ sudo ./setup-rocm.sh
Submodule 'dlas' (https://git.ecker.tech/mrq/DL-Art-School) registered for path 'dlas'
Submodule 'tortoise-tts' (https://git.ecker.tech/mrq/tortoise-tts) registered for path 'tortoise-tts'
Cloning into '/home/user/ai-voice-cloning/dlas'...
Cloning into '/home/user/ai-voice-cloning/tortoise-tts'...
Submodule path 'dlas': checked out '6c284ef8ec4c4769de3181d90ac96ff63581ef55'
Submodule path 'tortoise-tts': checked out '17b375c09d573315a0e5cda52a12e16ab5e48042'
./setup-rocm.sh: line 5: python: command not found
./setup-rocm.sh: line 6: ./venv/bin/activate: No such file or directory
./setup-rocm.sh: line 7: python: command not found
./setup-rocm.sh: line 9: pip: command not found
./setup-rocm.sh: line 10: python: command not found
./setup-rocm.sh: line 11: python: command not found
./setup-rocm.sh: line 12: python: command not found
./setup-rocm.sh: line 13: python: command not found
./setup-rocm.sh: line 14: deactivate: command not found

Is the script broken or is my system broken?
running this in terminal returns:

~/ai-voice-cloning$ python3 --version
Python 3.10.6

Many thanks for all your hard work.
MCS

Hello, me again. I've got Ubuntu set up now, got ROCM installed and AFAIK working, but I get command not found errors when attempting to run the install script: ``` > ~/ai-voice-cloning$ sudo ./setup-rocm.sh Submodule 'dlas' (https://git.ecker.tech/mrq/DL-Art-School) registered for path 'dlas' Submodule 'tortoise-tts' (https://git.ecker.tech/mrq/tortoise-tts) registered for path 'tortoise-tts' Cloning into '/home/user/ai-voice-cloning/dlas'... Cloning into '/home/user/ai-voice-cloning/tortoise-tts'... Submodule path 'dlas': checked out '6c284ef8ec4c4769de3181d90ac96ff63581ef55' Submodule path 'tortoise-tts': checked out '17b375c09d573315a0e5cda52a12e16ab5e48042' ./setup-rocm.sh: line 5: python: command not found ./setup-rocm.sh: line 6: ./venv/bin/activate: No such file or directory ./setup-rocm.sh: line 7: python: command not found ./setup-rocm.sh: line 9: pip: command not found ./setup-rocm.sh: line 10: python: command not found ./setup-rocm.sh: line 11: python: command not found ./setup-rocm.sh: line 12: python: command not found ./setup-rocm.sh: line 13: python: command not found ./setup-rocm.sh: line 14: deactivate: command not found ``` Is the script broken or is my system broken? running this in terminal returns: ``` ~/ai-voice-cloning$ python3 --version Python 3.10.6 ``` Many thanks for all your hard work. MCS

mrq commented

2023-02-21 20:19:41 +00:00

Right, some installs will not have python as an alias. I'll push a fix out, although I swore I did this already before, but it might have gotten reverted during migration from the old repo.

Right, some installs will not have `python` as an alias. I'll push a fix out, although I swore I did this already before, but it might have gotten reverted during migration from the old repo.

mrq commented

2023-02-21 20:21:23 +00:00

Should be remedied in commit f119993fb5.

Should be remedied in commit f119993fb5724dafe0e68cf0da360f3ab6c2f61d.

mrchucksneed commented

2023-02-21 20:36:18 +00:00

Should be remedied in commit f119993fb5.

Can confirm remedied and holy moly is ROCm way faster.
Thank you so much.

> Should be remedied in commit f119993fb5724dafe0e68cf0da360f3ab6c2f61d. Can confirm remedied and holy moly is ROCm way faster. Thank you so much.

mrq commented

2023-02-21 20:41:21 +00:00

Glad to hear.

holy moly is ROCm way faster

Yeah, DirectML unfortunately is still in its infancy, and I need to do some patches where it'll do some calculations on CPU due to some unimplemented hardware-side functions. Even ignoring those, it still consumes a lot of VRAM and already performs slower compared to CUDA (my 2060 at batch size 1 generates faster than my 6800XT at batch size 16). But it works, and that's good enough for users who have no choice (or absolutely refuse to use Linux).

Glad to hear. > holy moly is ROCm way faster Yeah, DirectML unfortunately is still in its infancy, and I need to do some patches where it'll do some calculations on CPU due to some unimplemented hardware-side functions. Even ignoring those, it still consumes a lot of VRAM and already performs slower compared to CUDA (my 2060 at batch size 1 generates faster than my 6800XT at batch size 16). But it works, and that's good enough for users who have no choice (or absolutely refuse to use Linux).

mrq closed this issue

2023-02-21 20:41:21 +00:00

Sign in to join this conversation.