pszemraj

import logging
import random
import time
from pathlib import Path

import fire
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer

logging.basicConfig(format="%(levelname)s - %(message)s", level=logging.INFO)

pszemraj

"""
The script is designed to monitor a specified directory for any file system changes (like additions, deletions, or modifications of files and subdirectories) and automatically upload the changes to a specified repository on the Hugging Face Hub.


pip install huggingface-hub watchdog
"""
import argparse
import logging
import time
from pathlib import Path

pszemraj

import os
import random

from datasets import load_dataset


def format_dataset(example):
    """Formats the dataset example into a single 'text' field."""

    # Add input only if it is longer than 2 characters

pszemraj

import logging
import subprocess

import torch


def check_ampere_gpu():
    """Check if the GPU supports NVIDIA Ampere or later and enable FP32 in PyTorch if it does."""

    # Check if CUDA is available

pszemraj

import os
import time

import requests


class Timer:
    """Basic timer utility."""

    def __enter__(self):

pszemraj

Useful misc installs

Details

Kernel and Low-Level Tools

1. Microcode Update: Keeping your CPU microcode updated can help in better performance and security. You can install the AMD microcode package by running:

sudo apt install amd64-microcode

pszemraj

"""
inference with a model trained on query well-formedness

https://huggingface.co/Ashishkr/query_wellformedness_score

pip transformers install accelerate optimum -q
"""
import torch
from transformers import AutoTokenizer, AutoModelForSequenceClassification

pszemraj

"""
run_langchain_summarization.py - Generate summaries using langchain + LLMs

For usage details, run `python run_langchain_summarization.py --help` and fire will print the usage details.

Notes:
- you need to have OPENAI_API_KEY set as an environment variable (easiest way is export OPENAI_API_KEY=memes123)
- install the dependencies using the requirements.txt file or below

    pip install fire langchain clean-text tqdm tiktoken

pszemraj

BEE-spoke-data/smol_llama-101M-GQA with Model Revision: .
Output dir: 101mgqa
Batch Size: 64
Device ID: cuda:2
Setting number of workers to 4
[dynet] random seed: 1234
[dynet] allocating memory: 32MB
[dynet] memory allocation done.
Token indices sequence length is longer than the specified maximum sequence length for this model (1377 > 1024). Running this sequence through the model will result in indexing errors
Selected Tasks: ['arc_easy', 'boolq', 'lambada_openai', 'openbookqa', 'piqa', 'winogrande']

pszemraj

from pathlib import Path
import logging
from typing import Optional, Union
import torch
from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline


# Function to process audio using distil-whisper
def process_audio_distil_whisper(
    audio_path: Union[str, Path],