pszemraj

LLM Foundry Configuration Reference

This document provides a comprehensive reference for all configuration options available in LLM Foundry YAML files. Configuration files are used for training, fine-tuning, and evaluating large language models.

Table of Contents

• Configuration Format Types
• Variables
• Model Configuration
• MPT Model Options

pszemraj

# -*- coding: utf-8 -*-
"""gemma-3n-test

pip install -U -q git+https://github.com/huggingface/transformers.git
pip install -U -q git+https://github.com/huggingface/pytorch-image-models.git
"""

from transformers import pipeline
import torch

pszemraj

#!/usr/bin/env python3
"""
Slice a (possibly very tall) image into fixed-height chunks.

Creates a sibling directory called  <image stem>_slices/
and writes   slice_000.png, slice_001.png, …  inside it.
"""

import argparse
from pathlib import Path

pszemraj

"""
Create & save an hf dataset with train/test/val splits from dir w/ text files

Ideal structure:

    root / section_name_1 / file 1
    root / section_name_1 / file 2
    root / section_name_1 / file YYY
    root / section_name_2 / file 1
    root / section_name_2 / file ZZZ

pszemraj

#!/usr/bin/env python
# -*- coding: utf-8 -*-
"""
Standalone Asynchronous Nanonets-OCR-s Inference Script using vLLM and PyMuPDF.

This script processes PDF files from an input directory using the
nanonets/Nanonets-OCR-s model served locally by vLLM via its OpenAI-compatible API.
It renders each page, sends API requests concurrently for OCR, extracts the
structured markdown/HTML text, and saves the combined text for each PDF into a
corresponding .txt file in the specified output directory.

pszemraj

from dataclasses import dataclass
from typing import List, Optional, Tuple

import torch
import torch.nn as nn


@dataclass
class _LayerSummary:
    """A dataclass to hold summary information for a single layer."""

pszemraj

"""
WaveNet: An Ultra-Small Language Model (PyTorch Implementation)

Based on the paper: https://arxiv.org/abs/2411.02674
Hugging Face Transformers compatible implementation.
"""
import math
from typing import Dict, Optional, Tuple, Union

import torch

pszemraj

The Enshittification of Closed-Weight Frontier Models

Ruminations on Theory and Motivations

1. The Concept of Enshittification: Coined by Cory Doctorow, it describes the pattern where platforms initially offer great value to attract users, then lock them in, and finally extract value by degrading the service for users while increasing value extraction for business customers (advertisers, etc.) or, in this case, the platform owner themselves by reducing costs.

2. Applying it to Frontier AI Chatbots:

   • Phase 1: Attract Users: Release a groundbreaking model (e.g., initial GPT-4, Claude 3 Opus). Offer free access or affordable subscriptions. Generate massive hype and positive press. Users are amazed by the capabilities (complex reasoning, creativity, coding).

• Phase 2: Lock-in Users: Users integrate the tool into their daily workflows, studies, or creative processes. They become accustomed to its abilities and interface. Subscription models create a direct financial lock-in

pszemraj

from typing import Dict, List, Optional

import datasets

# Optional: Keep the version print outside the function if desired
# print(f"Using datasets library version: {datasets.__version__}")


def create_interleaved_streaming_dataset(
    dataset_path: str = "Zyphra/Zyda-2",

pszemraj

#!/usr/bin/env python
"""
create_unified_mcqa.py  – “batteries‑included” multiple‑choice aggregator
✅ Handles all datasets listed in the conversation
✅ Survives missing/renamed columns
✅ Converts every `label` to pure int64 to avoid ClassLabel clashes
✅ Explicitly casts features to ensure concatenation compatibility
✅ Improved error handling and skipping for malformed examples
✅ Limits warning/info messages per dataset
✅ Fixes column mismatch error during cast