• Stars
    star
    293
  • Rank 136,961 (Top 3 %)
  • Language
    Jupyter Notebook
  • License
    Apache License 2.0
  • Created over 1 year ago
  • Updated about 2 months ago

Reviews

There are no reviews yet. Be the first to send feedback to the community and the maintainers!

Repository Details

Scaling Data-Constrained Language Models

Scaling Data-Constrained Language Models

This repository provides an overview of all components from the paper Scaling Data-Constrained Language Models.

We investigate scaling language models in data-constrained regimes. We run a large set of experiments varying the extent of data repetition and compute budget, ranging up to 900 billion training tokens and 9 billion parameter models. Based on our runs we propose and empirically validate a scaling law for compute optimality that accounts for the decreasing value of repeated tokens and excess parameters. We also experiment with approaches mitigating data scarcity, including augmenting the training dataset with code data, perplexity-filtering and deduplication. Models and datasets from our 400 training runs are available via this repository.

Data

Repeating

We experiment with repeating data on C4 and the non-deduplicated English split of OSCAR. For each dataset, we download the data and turn it into a single jsonl file, c4.jsonl and oscar_en.jsonl respectively.

Then we decide on the amount of unique tokens and the respective number of samples we need from the dataset. Note that C4 has 478.625834583 tokens per sample and OSCAR has 1312.0951072 with the GPT2Tokenizer. This was calculated by tokenizing the entire dataset and dividing the number of tokens by the number of samples. We use these numbers to calculate the needed samples.

For example, for 1.9B unique tokens, we need 1.9B / 478.625834583 = 3969697.96178 samples for C4 and 1.9B / 1312.0951072 = 1448065.76107 samples for OSCAR. To tokenize the data, we first need to clone the Megatron-DeepSpeed repository and follow its setup guide. We then select these samples and tokenize them as follows:

C4:

head -n 3969698 c4.jsonl > c4_1b9.jsonl
python Megatron-DeepSpeed/tools/preprocess_data_many_cores.py \
--input c4_1b9.jsonl \
--output-prefix gpt2tok_c4_en_1B9 \
--dataset-impl mmap \
--tokenizer-type PretrainedFromHF \
--tokenizer-name-or-path gpt2 \
--append-eod \
--workers 64

OSCAR:

head -n  1448066 oscar_en.jsonl > oscar_1b9.jsonl
python Megatron-DeepSpeed/tools/preprocess_data_many_cores.py \
--input oscar_1b9.jsonl \
--output-prefix gpt2tok_oscar_en_1B9 \
--dataset-impl mmap \
--tokenizer-type PretrainedFromHF \
--tokenizer-name-or-path gpt2 \
--append-eod \
--workers 64

where gpt2 points to a folder containing all files from https://huggingface.co/gpt2/tree/main. By using head we make sure that different subsets will have overlapping samples to reduce randomness.

For evaluation during training and the final evaluation, we use the validation set for C4:

from datasets import load_dataset
load_dataset("c4", "en", split="validation").to_json("c4-en-validation.json")
python Megatron-DeepSpeed/tools/preprocess_data_many_cores.py \
--input c4-en-validation.jsonl \
--output-prefix gpt2tok_c4validation_rerun \
--dataset-impl mmap \
--tokenizer-type PretrainedFromHF \
--tokenizer-name-or-path gpt2 \
--append-eod \
--workers 2

For OSCAR which has no official validation set we take a part of the training set by doing tail -364608 oscar_en.jsonl > oscarvalidation.jsonl and then tokenize it as follows:

python Megatron-DeepSpeed/tools/preprocess_data_many_cores.py --input oscarvalidation.jsonl --output-prefix gpt2tok_oscarvalidation --dataset-impl mmap --tokenizer-type PretrainedFromHF --tokenizer-name-or-path gpt2 --append-eod --workers 2

We have uploaded several preprocessed subsets for usage with megatron:

Some bin files were too large for git and thus split using e.g. split --number=l/40 gpt2tok_c4_en_1B9.bin gpt2tok_c4_en_1B9.bin. and split --number=l/40 gpt2tok_oscar_en_1B9.bin gpt2tok_oscar_en_1B9.bin.. To use them for training you need to cat them together again using cat gpt2tok_c4_en_1B9.bin.* > gpt2tok_c4_en_1B9.bin and cat gpt2tok_oscar_en_1B9.bin.* > gpt2tok_oscar_en_1B9.bin.

Code

We experiment with mixing code with the natural language data using the Python split from the the-stack-dedup. We download the data, turn it into a single jsonl file and preprocess it using the same approach as outlined above.

We have uploaded the preprocessed version for usage with megatron here: https://huggingface.co/datasets/datablations/python-megatron. We have split the bin file using split --number=l/40 gpt2tok_python_content_document.bin gpt2tok_python_content_document.bin., so you need to cat them together again using cat gpt2tok_python_content_document.bin.* > gpt2tok_python_content_document.bin for training.

Filtering

We create versions of C4 and OSCAR with perplexity and deduplication-related filtering metadata:

To recreate these metadata datasets there are instructions at filtering/README.md.

We provide the tokenized versions that can be used for training with Megatron at:

.bin files were split using something like split --number=l/10 gpt2tok_oscar_en_perplexity_25_text_document.bin gpt2tok_oscar_en_perplexity_25_text_document.bin., so you need to concatenate them back together via cat gpt2tok_oscar_en_perplexity_25_text_document.bin. > gpt2tok_oscar_en_perplexity_25_text_document.bin.

To recreate the tokenized versions given the metadata dataset,

  • OSCAR:
    • Deduplication: See filtering/deduplication/filter_oscar_jsonl.py
    • Perplexity: See below.
  • C4:
    • Deduplication: See below.
    • Perplexity: See below.

Perplexity

To create the perplexity percentiles, follow the below instructions.

C4:

from datasets import load_dataset
import numpy as np
ds = load_dataset("datablations/c4-filter", streaming=False, num_proc=128)

p_25 = np.percentile(ds["train"]["perplexity"], 25)
p_50 = np.percentile(ds["train"]["perplexity"], 50)
p_75 = np.percentile(ds["train"]["perplexity"], 75)

# 25 - 75th percentile
ds["train"].filter(lambda x: p_25 < x["perplexity"] < p_75, num_proc=128).to_json("c4_perplexty2575.jsonl", num_proc=128, force_ascii=False)
# 25th percentile
ds["train"].filter(lambda x: x["perplexity"] < p_25, num_proc=128).to_json("c4_perplexty25.jsonl", num_proc=128, force_ascii=False)
# 50th percentile
ds["train"].filter(lambda x: x["perplexity"] < p_50, num_proc=128).to_json("c4_perplexty50.jsonl", num_proc=128, force_ascii=False)

OSCAR:

from datasets import load_dataset
import numpy as np
ds = load_dataset("datablations/oscar-filter", use_auth_token=True, streaming=False, num_proc=128)

p_25 = np.percentile(ds["train"]["perplexity_score"], 25)
p_50 = np.percentile(ds["train"]["perplexity_score"], 50)

# 25th percentile
ds["train"].filter(lambda x: x["perplexity_score"] < p_25, num_proc=128).remove_columns(['meta', 'perplexity_score', 'text_length', 'url', 'domain', 'dup_ratio', 'pairs', 'repetitions', 'included_in_dedup', 'cluster', 'id']).to_json("oscar_perplexity25.jsonl", num_proc=128, force_ascii=False)
# 50th percentile
ds["train"].filter(lambda x: x["perplexity_score"] < p_50, num_proc=128).remove_columns(['meta', 'perplexity_score', 'text_length', 'url', 'domain', 'dup_ratio', 'pairs', 'repetitions', 'included_in_dedup', 'cluster', 'id']).to_json("oscar_perplexity50.jsonl", num_proc=128, force_ascii=False)

You can then tokenize the resulting jsonl files for training with Megatron as described in the Repeating section.

Deduplication

C4: For C4 you just need to remove all samples where the repetitions field is populated, via e.g.

from datasets import load_dataset
import numpy as np
ds = load_dataset("datablations/c4-dedup", use_auth_token=True, streaming=False, num_proc=128)
ds.filter(lambda x: not(x["repetitions"]).to_json('c4_dedup.jsonl', num_proc=128, force_ascii=False)

OSCAR: For OSCAR we provide a script at filtering/filter_oscar_jsonl.py to create the deduplicated dataset given the dataset with filtering metadata.

You can then tokenize the resulting jsonl files for training with Megatron as described in the Repeating section.

Models

Download

All models can be downloaded at https://huggingface.co/datablations.

Models are generally named as follows: lm1-{parameters}-{tokens}-{unique_tokens}, specifically individual models in the folders are named as: {parameters}{tokens}{unique_tokens}{optional specifier}, for example 1b12b8100m would be 1.1 billion params, 2.8 billion tokens, 100 million unique tokens. The xby (1b1, 2b8 etc.) convention introduces some ambiguity whether numbers belong to parameters or tokens, but you can always check the sbatch script in the respective folder to see the exact parameters / tokens / unique tokens.

The easiest way to download a single model is e.g.:

GIT_LFS_SKIP_SMUDGE=1 git clone https://huggingface.co/datablations/lm1-misc
cd lm1-misc; git lfs pull --include 146m14b400m/global_step21553

If this takes too long, you can also use wget to directly download individual files from the folder, e.g.:

wget https://huggingface.co/datablations/lm1-misc/resolve/main/146m14b400m/global_step21553/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt

For models corresponding to the experiments in the paper, consult the following repositories:

Other models not analysed in the paper:

Training

Regular models

We train models with our fork of Megatron-DeepSpeed that works with AMD GPUs (via ROCm): https://github.com/TurkuNLP/Megatron-DeepSpeed If you would like to use NVIDIA GPUs (via cuda), you can use the original library: https://github.com/bigscience-workshop/Megatron-DeepSpeed

You need to follow the setup instructions of either repository to create your environment (Our setup specific to LUMI is detailed in training/megdssetup.md).

Each model folder contains an sbatch script that was used to train the model. You can use these as a reference to train your own models adapting the necessary environment variables. The sbatch scripts reference some additional files:

  • *txt files that specify the data paths. You can find them at utils/datapaths/*, however, you will likely need to adapt the path to point to your dataset.
  • model_params.sh, which is at utils/model_params.sh and contains architecture presets.
  • launch.sh that you can find at training/launch.sh. It contains commands specific to our setup, which you may want to remove.

After training you can convert your model to transformers with e.g. python Megatron-DeepSpeed/tools/convert_checkpoint/deepspeed_to_transformers.py --input_folder global_step52452 --output_folder transformers --target_tp 1 --target_pp 1.

For repeat models, we also upload their tensorboards after training using e.g. tensorboard dev upload --logdir tensorboard_8b7178b88boscar --name "tensorboard_8b7178b88boscar", which makes them easy to use for visualization in the paper.

muP

For the muP ablation in the Appendix we use the script at training_scripts/mup.py. It contains setup instructions.

Parametric Fit

You can use our formula to compute the expected loss given parameters, data and unique tokens as follows:

import numpy as np
func = r"$L(N,D,R_N,R_D)=E + \frac{A}{(U_N + U_N * R_N^* * (1 - e^{(-1*R_N/(R_N^*))}))^\alpha} + \frac{B}{(U_D + U_D * R_D^* * (1 - e^{(-1*R_D/(R_D^*))}))^\beta}$"
a, b, e, alpha, beta, rd_star, rn_star = [6.255414, 7.3049974, 0.6254804, 0.3526596, 0.3526596, 15.387756, 5.309743]

A = np.exp(a)
B = np.exp(b)
E = np.exp(e)
G = ((alpha*A)/(beta*B))**(1/(alpha+beta))

def D_to_N(D):
    return (D * G)**(beta/alpha) * G

def scaling_law(N, D, U):
    """
    N: number of parameters
    D: number of total training tokens
    U: number of unique training tokens
    """
    assert U <= D, "Cannot have more unique tokens than total tokens"

    RD = np.maximum((D / U) - 1, 0)    
    UN = np.minimum(N, D_to_N(U))
    RN = np.maximum((N / UN ) - 1, 0)

    L = E + A/(UN + UN*rn_star*(1-np.exp(-1*RN/rn_star)))**alpha + B / (U + U * rd_star * (1 - np.exp(-1*RD/(rd_star))))**beta
    return L

# Models in Figure 1 (right):
print(scaling_law(6.34e9, 242e9, 25e9)) # 2.2256440889984477 # <- This one is better
print(scaling_law(8.67e9, 178e9, 25e9)) # 2.2269634075087867

Note that the actual loss value is unlikely to be useful, but rather the trend of the loss as e.g. the number of parameters increases or to compare two models like in the example above. To compute the optimal allocation, you can use a simple grid search:

def optimal_allocation(N_BASE, D_BASE, U_BASE):
    min_l = float("inf")
    for i in np.linspace(1.0001, 3, 500):
        D =  D_BASE*i
        U = min(U_BASE, D)
        N = N_BASE/i
        new_l = scaling_law(N, D, U)
        if new_l < min_l:
            min_l, min_t, min_s = new_l, D, N
        D =  D_BASE/i
        U = min(U_BASE, D)
        N = N_BASE*i
        new_l = scaling_law(N, D, U)
        if new_l < min_l:
            min_l, min_t, min_s = new_l, D, N
     return min_l, min_t, min_s

_, min_t, min_s = optimal_allocation(8.67e9, 178e9, 25e9)
print(f"Optimal configuration: {min_t} tokens, {min_t/25e9} epochs, {min_s} parameters")
# -> 227B tokens, 9.1 epochs, 6.8B parameters
# We went more extreme in Figure 1 to really put our prediction of "many epochs, fewer params" to the test

If you derive a closed-form expression for the optimal allocation instead of the above grid search, please let us know :) We fit data-constrained scaling laws & the C4 scaling coefficients using the code at utils/parametric_fit.ipynb equivalent to this colab.

Downstream Evaluation

Rank Eval / Accuracy

  1. Follow the instructions in the Training > Regular models section to setup a training environment.
  2. Install the evaluation harness: pip install git+https://github.com/EleutherAI/lm-evaluation-harness.git. We used version 0.2.0, but newer versions should work as well.
  3. Run sbatch utils/eval_rank.sh modifying the necessary variables in the script first
  4. After running, we convert each file to csv using python Megatron-DeepSpeed/tasks/eval_harness/report-to-csv.py outfile.json

Generative / Rouge

  1. Clone the addtasks branch of the evaluation harness: git clone -b addtasks https://github.com/Muennighoff/lm-evaluation-harness.git
  2. Setup an environment with cd lm-evaluation-harness; pip install -e ".[dev]"; pip uninstall -y promptsource; pip install git+https://github.com/Muennighoff/promptsource.git@tr13 i.e. all requirements except promptsource, which is installed from a fork with the correct prompts
  3. Make sure your checkpoint path is a transformers checkpoint path
  4. Run sbatch utils/eval_generative.sh modifying the necessary variables in the script first
  5. After running, we merge the generation files using python utils/merge_generative.py and then convert them to csv with python utils/csv_generative.py merged.json

bAbI / Exact match

  1. Clone the babi branch of the evaluation harness: git clone -b babi https://github.com/Muennighoff/lm-evaluation-harness.git (Note that this branch is not compatible with the addtasks branch for generative tasks as it stems from EleutherAI/lm-evaluation-harness, while addtasks is based on bigscience/lm-evaluation-harness)
  2. Setup an environment with cd lm-evaluation-harness; pip install -e ".[dev]"
  3. Make sure your checkpoint path is a transformers checkpoint with tokenizer files (If you trained a gpt2 model like all models in this work, it's just the files from here: https://huggingface.co/gpt2)
  4. Run sbatch utils/eval_babi.sh modifying the necessary variables in the script first

Plots & Tables

Plots

  • Figure 1: plotstables/return_alloc.pdf, plotstables/return_alloc.ipynb, colab
  • Figure 2: plotstables/dataset_setup.pdf, plotstables/dataset_setup.ipynb, colab
  • Figure 3: plotstables/contours.pdf, plotstables/contours.ipynb, colab
  • Figure 4: plotstables/isoflops_training.pdf, plotstables/isoflops_training.ipynb, colab
  • Figure 5: plotstables/return.pdf, plotstables/return.ipynb, colab
  • Figure 6 (Left): plotstables/strategies.pdf, plotstables/strategies.drawio
  • Figure 6 (Right): plotstables/beyond.pdf, plotstables/beyond.ipynb, colab
  • Figure 7: plotstables/cartoon.pdf, plotstables/cartoon.pptx
  • Figure 8: plotstables/isoloss_400m1b5.pdf & same colab as Figure 3
  • Figure 9 - 11: plotstables/mup.pdf, plotstables/dd.pdf, plotstables/dedup.pdf, plotstables/mup_dd_dd.ipynb, colab
  • Figure 12: plotstables/isoloss_alphabeta_100m.pdf & same colab as Figure 3
  • Figure 13: plotstables/galactica.pdf, plotstables/galactica.ipynb, colab
  • Figure 14 - 17: training_c4.pdf, validation_c4oscar.pdf, training_oscar.pdf, validation_epochs_c4oscar.pdf & same colab as Figure 4
  • Figure 18: plotstables/perplexity_histogram.pdf, plotstables/perplexity_histogram.ipynb
  • Figure 19 - 20: plotstabls/validation_c4py.pdf, plotstables/training_validation_filter.pdf, plotstables/beyond_losses.ipynb & colab
  • Figure 21 - 39: Manual

Tables

  • Table 1-2: Manual
  • Table 3 - 8: plotstables/repetition.ipynb & colab
  • Table 9 - 10: plotstables/python.ipynb & colab
  • Table 11: Manual
  • Table 12 - 13: plotstables/filtering.ipynb & colab
  • Table 14: Manual

License

All models & code are licensed under Apache 2.0. Filtered datasets are released with the same license as the datasets they stem from.

Citation

@article{muennighoff2023scaling,
  title={Scaling Data-Constrained Language Models},
  author={Muennighoff, Niklas and Rush, Alexander M and Barak, Boaz and Scao, Teven Le and Piktus, Aleksandra and Tazi, Nouamane and Pyysalo, Sampo and Wolf, Thomas and Raffel, Colin},
  journal={arXiv preprint arXiv:2305.16264},
  year={2023}
}

More Repositories

1

transformers

πŸ€— Transformers: State-of-the-art Machine Learning for Pytorch, TensorFlow, and JAX.
Python
125,891
star
2

pytorch-image-models

PyTorch image models, scripts, pretrained weights -- ResNet, ResNeXT, EfficientNet, NFNet, Vision Transformer (ViT), MobileNet-V3/V2, RegNet, DPN, CSPNet, Swin Transformer, MaxViT, CoAtNet, ConvNeXt, and more
Python
28,073
star
3

diffusers

πŸ€— Diffusers: State-of-the-art diffusion models for image and audio generation in PyTorch and FLAX.
Python
22,776
star
4

datasets

πŸ€— The largest hub of ready-to-use datasets for ML models with fast, easy-to-use and efficient data manipulation tools
Python
17,530
star
5

peft

πŸ€— PEFT: State-of-the-art Parameter-Efficient Fine-Tuning.
Python
14,007
star
6

candle

Minimalist ML framework for Rust
Rust
12,686
star
7

tokenizers

πŸ’₯ Fast State-of-the-Art Tokenizers optimized for Research and Production
Rust
8,286
star
8

trl

Train transformer language models with reinforcement learning.
Python
8,181
star
9

text-generation-inference

Large Language Model Text Generation Inference
Python
7,240
star
10

accelerate

πŸš€ A simple way to launch, train, and use PyTorch models on almost any device and distributed configuration, automatic mixed precision (including fp8), and easy-to-configure FSDP and DeepSpeed support
Python
7,008
star
11

chat-ui

Open source codebase powering the HuggingChat app
TypeScript
6,369
star
12

deep-rl-class

This repo contains the syllabus of the Hugging Face Deep Reinforcement Learning Course.
MDX
3,541
star
13

alignment-handbook

Robust recipes to align language models with human and AI preferences
Python
3,485
star
14

autotrain-advanced

πŸ€— AutoTrain Advanced
Python
3,283
star
15

diffusion-models-class

Materials for the Hugging Face Diffusion Models Course
Jupyter Notebook
3,126
star
16

notebooks

Notebooks using the Hugging Face libraries πŸ€—
Jupyter Notebook
3,114
star
17

distil-whisper

Distilled variant of Whisper for speech recognition. 6x faster, 50% smaller, within 1% word error rate.
Python
2,964
star
18

neuralcoref

✨Fast Coreference Resolution in spaCy with Neural Networks
C
2,806
star
19

knockknock

πŸšͺ✊Knock Knock: Get notified when your training ends with only two additional lines of code
Python
2,682
star
20

swift-coreml-diffusers

Swift app demonstrating Core ML Stable Diffusion
Swift
2,377
star
21

safetensors

Simple, safe way to store and distribute tensors
Python
2,347
star
22

optimum

πŸš€ Accelerate training and inference of πŸ€— Transformers and πŸ€— Diffusers with easy to use hardware optimization tools
Python
2,086
star
23

awesome-papers

Papers & presentation materials from Hugging Face's internal science day
1,996
star
24

blog

Public repo for HF blog posts
Jupyter Notebook
1,962
star
25

setfit

Efficient few-shot learning with Sentence Transformers
Jupyter Notebook
1,912
star
26

text-embeddings-inference

A blazing fast inference solution for text embeddings models
Rust
1,845
star
27

course

The Hugging Face course on Transformers
MDX
1,832
star
28

evaluate

πŸ€— Evaluate: A library for easily evaluating machine learning models and datasets.
Python
1,825
star
29

transfer-learning-conv-ai

πŸ¦„ State-of-the-Art Conversational AI with Transfer Learning
Python
1,654
star
30

swift-coreml-transformers

Swift Core ML 3 implementations of GPT-2, DistilGPT-2, BERT, and DistilBERT for Question answering. Other Transformers coming soon!
Swift
1,543
star
31

pytorch-openai-transformer-lm

πŸ₯A PyTorch implementation of OpenAI's finetuned transformer language model with a script to import the weights pre-trained by OpenAI
Python
1,464
star
32

cookbook

Open-source AI cookbook
Jupyter Notebook
1,357
star
33

huggingface_hub

All the open source things related to the Hugging Face Hub.
Python
1,311
star
34

Mongoku

πŸ”₯The Web-scale GUI for MongoDB
TypeScript
1,289
star
35

huggingface.js

Utilities to use the Hugging Face Hub API
TypeScript
1,193
star
36

hmtl

🌊HMTL: Hierarchical Multi-Task Learning - A State-of-the-Art neural network model for several NLP tasks based on PyTorch and AllenNLP
Python
1,185
star
37

gsplat.js

JavaScript Gaussian Splatting library.
TypeScript
1,114
star
38

llm-vscode

LLM powered development for VSCode
TypeScript
1,060
star
39

datatrove

Freeing data processing from scripting madness by providing a set of platform-agnostic customizable pipeline processing blocks.
Python
1,033
star
40

pytorch-pretrained-BigGAN

πŸ¦‹A PyTorch implementation of BigGAN with pretrained weights and conversion scripts.
Python
986
star
41

torchMoji

πŸ˜‡A pyTorch implementation of the DeepMoji model: state-of-the-art deep learning model for analyzing sentiment, emotion, sarcasm etc
Python
880
star
42

nanotron

Minimalistic large language model 3D-parallelism training
Python
810
star
43

naacl_transfer_learning_tutorial

Repository of code for the tutorial on Transfer Learning in NLP held at NAACL 2019 in Minneapolis, MN, USA
Python
718
star
44

awesome-huggingface

πŸ€— A list of wonderful open-source projects & applications integrated with Hugging Face libraries.
698
star
45

optimum-nvidia

Python
680
star
46

dataset-viewer

Lightweight web API for visualizing and exploring any dataset - computer vision, speech, text, and tabular - stored on the Hugging Face Hub
Python
614
star
47

transformers-bloom-inference

Fast Inference Solutions for BLOOM
Python
546
star
48

exporters

Export Hugging Face models to Core ML and TensorFlow Lite
Python
540
star
49

pytorch_block_sparse

Fast Block Sparse Matrices for Pytorch
C++
523
star
50

llm.nvim

LLM powered development for Neovim
Lua
507
star
51

swift-transformers

Swift Package to implement a transformers-like API in Swift
Swift
482
star
52

node-question-answering

Fast and production-ready question answering in Node.js
TypeScript
459
star
53

large_language_model_training_playbook

An open collection of implementation tips, tricks and resources for training large language models
Python
431
star
54

llm-ls

LSP server leveraging LLMs for code completion (and more?)
Rust
416
star
55

llm_training_handbook

An open collection of methodologies to help with successful training of large language models.
Python
385
star
56

swift-chat

Mac app to demonstrate swift-transformers
Swift
375
star
57

tflite-android-transformers

DistilBERT / GPT-2 for on-device inference thanks to TensorFlow Lite with Android demo apps
Java
368
star
58

community-events

Place where folks can contribute to πŸ€— community events
Jupyter Notebook
368
star
59

nn_pruning

Prune a model while finetuning or training.
Jupyter Notebook
360
star
60

text-clustering

Easily embed, cluster and semantically label text datasets
Python
335
star
61

speechbox

Python
328
star
62

100-times-faster-nlp

πŸš€100 Times Faster Natural Language Processing in Python - iPython notebook
HTML
325
star
63

education-toolkit

Educational materials for universities
Jupyter Notebook
307
star
64

controlnet_aux

Python
306
star
65

optimum-intel

πŸ€— Optimum Intel: Accelerate inference with Intel optimization tools
Jupyter Notebook
295
star
66

unity-api

C#
284
star
67

open-muse

Open reproduction of MUSE for fast text2image generation.
Python
284
star
68

audio-transformers-course

The Hugging Face Course on Transformers for Audio
MDX
247
star
69

hub-docs

Docs of the Hugging Face Hub
221
star
70

lighteval

LightEval is a lightweight LLM evaluation suite that Hugging Face has been using internally with the recently released LLM data processing library datatrove and LLM training library nanotron.
Python
208
star
71

quanto

A pytorch Quantization Toolkit
Python
201
star
72

simulate

🎒 Creating and sharing simulation environments for embodied and synthetic data research
Python
185
star
73

ratchet

A cross-platform browser ML framework.
Rust
184
star
74

optimum-benchmark

A unified multi-backend utility for benchmarking Transformers, Timm, Diffusers and Sentence-Transformers with full support of Optimum's hardware optimizations & quantization schemes.
Python
183
star
75

hf_transfer

Rust
181
star
76

olm-datasets

Pipeline for pulling and processing online language model pretraining data from the web
Python
169
star
77

instruction-tuned-sd

Code for instruction-tuning Stable Diffusion.
Python
167
star
78

optimum-neuron

Easy, fast and very cheap training and inference on AWS Trainium and Inferentia chips.
Jupyter Notebook
163
star
79

llm-swarm

Manage scalable open LLM inference endpoints in Slurm clusters
Python
156
star
80

OBELICS

Code used for the creation of OBELICS, an open, massive and curated collection of interleaved image-text web documents, containing 141M documents, 115B text tokens and 353M images.
Python
147
star
81

workshops

Materials for workshops on the Hugging Face ecosystem
Jupyter Notebook
146
star
82

cosmopedia

Python
138
star
83

api-inference-community

Python
131
star
84

diffusion-fast

Faster generation with text-to-image diffusion models.
Python
127
star
85

diarizers

Python
106
star
86

optimum-habana

Easy and lightning fast training of πŸ€— Transformers on Habana Gaudi processor (HPU)
Python
106
star
87

sharp-transformers

A Unity plugin for using Transformers models in Unity.
C#
104
star
88

competitions

Python
101
star
89

hf-hub

Rust client for the huggingface hub aiming for minimal subset of features over `huggingface-hub` python package
Rust
93
star
90

olm-training

Repo for training MLMs, CLMs, or T5-type models on the OLM pretraining data, but it should work with any hugging face text dataset.
Python
87
star
91

fuego

[WIP] A πŸ”₯ interface for running code in the cloud
Python
84
star
92

tune

Python
83
star
93

datasets-viewer

Viewer for the πŸ€— datasets library.
Python
82
star
94

optimum-graphcore

Blazing fast training of πŸ€— Transformers on Graphcore IPUs
Python
78
star
95

frp

FRP Fork
Go
73
star
96

paper-style-guide

72
star
97

block_movement_pruning

Block Sparse movement pruning
Python
70
star
98

amused

Python
68
star
99

doc-builder

The package used to build the documentation of our Hugging Face repos
Python
67
star
100

data-measurements-tool

Developing tools to automatically analyze datasets
Python
67
star