willccbb

# train_grpo.py
#
# See https://github.com/willccbb/verifiers for ongoing developments
#
import re
import torch
from datasets import load_dataset, Dataset
from transformers import AutoTokenizer, AutoModelForCausalLM
from peft import LoraConfig
from trl import GRPOConfig, GRPOTrainer

veekaybee

Anti-hype LLM reading list

Goals: Add links that are reasonable and good explanations of how stuff works. No hype and no vendor content if possible. Practical first-hand accounts of models in prod eagerly sought.

Foundational Concepts

Pre-Transformer Models

Birch-san

Installing CUDA 12.1.1 + PyTorch nightly + Python 3.10 on Ubuntu 22.10

Should you keep your NVIDIA driver?

CUDA 12.1.1 toolkit is gonna offer to install Nvidia driver 530 for us. It's from New Feature branch. It's likely to be newer than the default Nvidia driver you would've installed via apt-get (apt would prefer to give you 525, i.e. Production Branch).

If you're confident that you already have a new enough Nvidia driver for CUDA 12.1.1, and you'd like to keep your driver: feel free to skip this "uninstall driver" step.

But if you're not sure, or you know your driver is too old: let's uninstall it. CUDA will install a new driver for us later.

yoavg

Reinforcement Learning for Language Models

Yoav Goldberg, April 2023.

Why RL?

With the release of the ChatGPT model and followup large language models (LLMs), there was a lot of discussion of the importance of "RLHF training", that is, "reinforcement learning from human feedback". I was puzzled for a while as to why RL (Reinforcement Learning) is better than learning from demonstrations (a.k.a supervised learning) for training language models. Shouldn't learning from demonstrations (or, in language model terminology "instruction fine tuning", learning to immitate human written answers) be sufficient? I came up with a theoretical argument that was somewhat convincing. But I came to realize there is an additional argumment which not only supports the case of RL training, but also requires it, in particular for models like ChatGPT. This additional argument is spelled out in (the first half of) a talk by John Schulman from OpenAI. This post pretty much

NaxAlpha

import time
from contextlib import suppress

import torch
import torch.nn as nn
import torch.optim as optim
import torch.nn.functional as F
import torch.backends.cuda as cuda
from torch.utils.data import DataLoader, IterableDataset

Helw150

from transformers import AutoTokenizer, T5ForConditionalGeneration

# Model Init
n_gpu = 8
tokenizer = AutoTokenizer.from_pretrained("google/flan-ul2")
model = T5ForConditionalGeneration.from_pretrained("google/flan-ul2")
heads_per_gpu = len(model.encoder.block) // n_gpu 
device_map = {
    gpu: list(
        range(

priyathamkat

import torch
import torch.multiprocessing as mp
from absl import app, flags
from torchvision.models import AlexNet

FLAGS = flags.FLAGS

flags.DEFINE_integer("num_processes", 2, "Number of subprocesses to use")

joshlk

NVIDIA GPU P2P Benchmark bandwidth/throughput and latency

Using https://github.com/NVIDIA/cuda-samples

You can also view the GPU topology using nvidia-smi topo -m

1. Download repo git clone https://github.com/NVIDIA/cuda-samples.git
2. Checkout the tag that corresponds with the right CUDA version: git checkout tags/v11.1
3. You might need to install some additional packages sudo apt-get install freeglut3-dev build-essential libx11-dev libxmu-dev libxi-dev libgl1-mesa-glx libglu1-mesa libglu1-mesa-dev libglfw3-dev libgles2-mesa-dev
4. Either build everything by just execting make in root dir. Or cd Samples/p2pBandwidthLatencyTest; make

growlnx

PROJECT(project C)

CMAKE_MINIMUM_REQUIRED(VERSION 3.1.0)

FIND_PACKAGE(PkgConfig REQUIRED)

PKG_CHECK_MODULES(GLIB REQUIRED glib-2.0)

INCLUDE_DIRECTORIES(
	src/include/

thomwolf

def top_k_top_p_filtering(logits, top_k=0, top_p=0.0, filter_value=-float('Inf')):
    """ Filter a distribution of logits using top-k and/or nucleus (top-p) filtering
        Args:
            logits: logits distribution shape (vocabulary size)
            top_k >0: keep only top k tokens with highest probability (top-k filtering).
            top_p >0.0: keep the top tokens with cumulative probability >= top_p (nucleus filtering).
                Nucleus filtering is described in Holtzman et al. (http://arxiv.org/abs/1904.09751)
    """
    assert logits.dim() == 1  # batch size 1 for now - could be updated for more but the code would be less clear
    top_k = min(top_k, logits.size(-1))  # Safety check