lmmx

conda create -n gptneox python numpy -y
conda activate gptneox
conda install pytorch torchvision pytorch-cuda=11.7 -c pytorch -c nvidia -y
pip install transformers accelerate bitsandbytes

lmmx

conda create -n coca python
conda activate coca
conda install pytorch torchvision pytorch-cuda=11.7 -c pytorch -c nvidia
conda install transformers open-clip-torch -c conda-forge

CPU version:

lmmx

class A:
    x = [1]

    @classmethod
    def fetch(cls, name: str = "x") -> list[int]:
        is_base_cls = super().__thisclass__ is cls
        parent_cls = cls.mro()[1]
        prev = [] if is_base_cls else parent_cls.fetch(name=name)
        return prev + getattr(cls, name, [])

lmmx

from enum import Enum

from pytest import fixture, mark


class Int1(Enum):
    small = 0
    large = 100

lmmx

    import ast

from pyflakes import checker, reporter


class CheckerSnooper(checker.Checker):
    verbose: bool

    def __init__(self, *args, **kwargs):
        self.verbose = kwargs.pop("verbose", False)

lmmx

lmmx

recfile="2022-10-01 19-44-24"; ffmpeg -i "$recfile.mkv" -q:a 0 -map a "$recfile.mp3"; whisper --model large --language en "$recfile.mp3"

lmmx

Louis "Hi Craig"

Craig "Hi Louis"

Louis "How much wood could a woodchuck chuck if a woodchuck could chuck wood?"

Craig "He would chuck, he would, as much as he could, and chuck as much wood as a woodchuck would if a woodchuck could chuck wood"

Louis "I just checked and in 1988, a New York fish and wildlife technician named Richard Thomas calculated the volume of dirt in a typical 25–30-foot long woodchuck burrow and determined that if the woodchuck had moved an equivalent volume of wood, it could move "about 700 pounds (320 kg) on a good day, with the wind at his back"."

lmmx

# ffmpeg -i foo.m4a foo.wav

from pyannote.audio import Pipeline

pipeline = Pipeline.from_pretrained("pyannote/speaker-diarization")
diarization = pipeline("foo.wav")

# RTTM format https://catalog.ldc.upenn.edu/docs/LDC2004T12/RTTM-format-v13.pdf
with open("foo.rttm", "w") as rttm:
    diarization.write_rttm(rttm)

lmmx

conda create -n diffusers
conda activate diffusers
conda install -y pytorch torchvision torchaudio cudatoolkit=11.6 -c pytorch -c conda-forge
pip install diffusers transformers piexif

OUTDATED For hf_stablediffusion_img2img.py download https://github.com/huggingface/diffusers/blob/main/examples/inference/image_to_image.py (edit: this has been deprecated)

For hf_stablediffusion_inpaint.py run pip install git+https://github.com/huggingface/diffusers.git (not yet been released). At time of writing the DDIM scheduler worked, the default PNDM one did not.