sroecker (Steffen Röcker)

upvoted a collection about 7 hours ago

Hermes

Collection

Nous' Flagship LLM Series • 21 items • Updated May 15 • 91

upvoted a paper about 15 hours ago

Replacing Judges with Juries: Evaluating LLM Generations with a Panel of Diverse Models

Paper • 2404.18796 • Published Apr 29 • 67

upvoted 2 collections 1 day ago

TabuLa-8B

Collection

Training, eval suite, and model from the paper "Large Scale Transfer Learning for Tabular Data via Language Modeling" https://arxiv.org/abs/2406.12031 • 4 items • Updated 2 days ago • 7

Depth Anything v2 Release

Collection

A comprehensive collection on DAv2 • 5 items • Updated 3 days ago • 8

upvoted 2 collections 2 days ago

Florence

Collection

9 items • Updated 6 days ago • 102

BM25S Indices

Collection

https://github.com/xhluca/bm25s • 14 items • Updated 2 days ago • 5

upvoted a collection 4 days ago

DeepSeekCoder-V2

Collection

4 items • Updated 7 days ago • 41

upvoted a collection 6 days ago

Small LLMs

Collection

Collection of Fine Tuned Small LLMs • 13 items • Updated 27 days ago • 2

upvoted a collection 7 days ago

FP8 LLMs for vLLM

Collection

Accurate FP8 quantized models by Neural Magic, ready for use with vLLM! • 10 items • Updated 1 day ago • 15

upvoted a paper 7 days ago

Magpie: Alignment Data Synthesis from Scratch by Prompting Aligned LLMs with Nothing

Paper • 2406.08464 • Published 8 days ago • 42

upvoted an article 7 days ago

Article

Putting RL back in RLHF

9 days ago

• 41

upvoted a collection 9 days ago

codestral-text2cypher

Collection

codestral finetuned for text2cypher • 3 items • Updated 11 days ago • 2

upvoted 2 collections 14 days ago

Local Function Calling Gems

Collection

These are the best function calling LLMs one can run on less than 64GB VRAM/Unified Memory. I use these on a M1 Max Macbook 64GB. • 5 items • Updated 15 days ago • 3

Qwen2

Collection

Qwen2 language models, including pretrained and instruction-tuned models of 5 sizes, including 0.5B, 1.5B, 7B, 57B-A14B, and 72B. • 29 items • Updated 15 days ago • 195

upvoted 2 collections 16 days ago

GLM-4

Collection

GLM-4 Open Models • 4 items • Updated 16 days ago • 76

DeTikZify

Collection

Synthesizing Graphics Programs for Scientific Figures and Sketches with TikZ • 9 items • Updated 18 days ago • 2

upvoted an article 17 days ago

Article

Uncensor any LLM with abliteration

By

•

8 days ago

• 221

upvoted an article 18 days ago

Article

Releasing Common Corpus: the largest public domain dataset for training LLMs

By

•

Mar 20

• 12

upvoted an article 21 days ago

Article

How to directly access 150k+ Hugging Face Datasets with DuckDB and query using GPT-4o

By

•

21 days ago

• 10

upvoted an article 22 days ago

Article

⚗️ 🔥 Building High-Quality Datasets with distilabel and Prometheus 2

By

•

17 days ago

• 20

upvoted a collection 22 days ago

sentence-transformers-from-synthetic-data

Collection

Example of using distilabel to generate synthetic triplets data for fine-tuning a Sentence Transformer model • 4 items • Updated 10 days ago • 18

upvoted an article 24 days ago

Article

Training and Finetuning Embedding Models with Sentence Transformers v3

24 days ago

• 100

upvoted a paper 24 days ago

Granite Code Models: A Family of Open Foundation Models for Code Intelligence

Paper • 2405.04324 • Published May 7 • 14

upvoted a collection 25 days ago

🤖Phi-3

Collection

6 items • Updated 18 days ago • 1

upvoted an article 27 days ago

Article

GPU Poor Savior: Revolutionizing Low-Bit Open Source LLMs and Cost-Effective Edge Computing

By

•

27 days ago

• 9

upvoted 2 collections 27 days ago

DiscoLeo 8B: Llama3 for German

Collection

Continued Pretraining on Llama3 8B to improve German linguistic capabilities. A collection of base and fine-tuned models and variants. • 5 items • Updated 27 days ago • 13

DiscoLeo 8B quants

Collection

A collection of different quantizations of the DiscoLeo models. • 3 items • Updated 27 days ago • 3

upvoted a collection 28 days ago

C4AI Aya 23

Collection

Aya 23 is an open weights research release of an instruction fine-tuned model with highly advanced multilingual capabilities. • 3 items • Updated 29 days ago • 34

upvoted an article 28 days ago

Article

⚗️ 🧑🏼‍🌾 Let's grow some Domain Specific Datasets together

By

•

Apr 29

• 27

upvoted a collection 29 days ago

C4AI Command R Plus

Collection

C4AI Command R+ is an open weights research release of a 104B billion parameter model with highly advanced capabilities. • 3 items • Updated 29 days ago • 20

upvoted an article 29 days ago

Article

Let's talk about LLM evaluation

By

•

29 days ago

• 92

upvoted 4 collections about 1 month ago

upvoted a paper about 2 months ago

What matters when building vision-language models?

Paper • 2405.02246 • Published May 3 • 89

upvoted 2 collections about 2 months ago

Granite Code Models

Collection

A series of code models trained by IBM licensed under Apache 2.0 license. We release both the base pretrained and instruct models. • 18 items • Updated 21 days ago • 145

SPPO

Collection

Self-Play Preference Optimization • 4 items • Updated May 4 • 2

upvoted 3 articles about 2 months ago

Article

Saving Memory Using Padding-Free Transformer Layers during Finetuning

By

•

10 days ago

• 6

Article

Introducing Idefics2: A Powerful 8B Vision-Language Model for the community

Apr 15

• 142

Article

🧑‍⚖️ "Replacing Judges with Juries" using distilabel

By

•

May 3

• 14

upvoted a collection about 2 months ago

llama 3 self-align experiments

Collection

Replicating the pipeline for StarCoder-2 Instruct on Llama-3-8B with some tweaks https://huggingface.co/blog/sc2-instruct • 4 items • Updated May 9 • 6

upvoted 2 articles about 2 months ago

Article

Improving Prompt Consistency with Structured Generations

Apr 30

• 48

Article

StarCoder2-Instruct: Fully Transparent and Permissive Self-Alignment for Code Generation

Apr 29

• 69

upvoted a collection about 2 months ago

LLaVA-Phi-3-mini

Collection

4 items • Updated Apr 28 • 11

upvoted 3 articles about 2 months ago

Article

Design choices for Vision Language Models in 2024

By

•

Apr 16

• 20

Article

Post-OCR-Correction: 1 billion words dataset of automated OCR correction by LLM

By

•

Apr 26

• 12

Article

LLM Comparison/Test: Llama 3 Instruct 70B + 8B HF/GGUF/EXL2 (20 versions tested and compared!)

By

•

Apr 24

• 51

upvoted a collection 2 months ago

〽️MistralAI

Collection

A collection of MistralAI models that you can trust in production! • 10 items • Updated 18 days ago • 7

upvoted a collection 5 months ago

Zeroshot Classifiers

Collection

These are my current best zeroshot classifiers. Some of my older models are downloaded more often, but the models in this collection are newer/better. • 11 items • Updated Apr 3 • 82

upvoted a collection 6 months ago

LLM Speculative Decoding

Collection

Tiny language models meant to serve as draft models for speculative decoding. • 6 items • Updated Jan 6 • 2

upvoted a collection 9 months ago

Recent models: last 100 repos, sorted by creation date

Collection

The last 100 repos I have created. Sorted by creation date descending, so the most recently created repos appear at the top. • 121 items • Updated Jan 31 • 455

Steffen Röcker PRO

AI & ML interests

Organizations

sroecker's activity

Putting RL back in RLHF

Uncensor any LLM with abliteration

Releasing Common Corpus: the largest public domain dataset for training LLMs

How to directly access 150k+ Hugging Face Datasets with DuckDB and query using GPT-4o

⚗️ 🔥 Building High-Quality Datasets with distilabel and Prometheus 2

Training and Finetuning Embedding Models with Sentence Transformers v3

GPU Poor Savior: Revolutionizing Low-Bit Open Source LLMs and Cost-Effective Edge Computing

⚗️ 🧑🏼‍🌾 Let's grow some Domain Specific Datasets together

Let's talk about LLM evaluation

Saving Memory Using Padding-Free Transformer Layers during Finetuning

Introducing Idefics2: A Powerful 8B Vision-Language Model for the community

🧑‍⚖️ "Replacing Judges with Juries" using distilabel

Improving Prompt Consistency with Structured Generations

StarCoder2-Instruct: Fully Transparent and Permissive Self-Alignment for Code Generation

Design choices for Vision Language Models in 2024

Post-OCR-Correction: 1 billion words dataset of automated OCR correction by LLM

LLM Comparison/Test: Llama 3 Instruct 70B + 8B HF/GGUF/EXL2 (20 versions tested and compared!)