Experience
-
-
-
-
-
-
-
-
-
Education
Honors & Awards
-
Konrad-Adenauer-Stipendium
Konrad-Adenauer-Striftung
-
Mannheim Stipendium
Universität Mannheim
Languages
-
Deutsch
Native or bilingual proficiency
-
Englisch
Full professional proficiency
Other similar profiles
Explore more posts
-
Eumentis
Learn how you can convert a PyTorch model, trained on the web, to a mobile-optimized format. Madhur Zanwar describes the process using a few lines of code. He also discusses his journey and how one approach didn't yield satisfactory results while trying to run it in a React Native app. #machinelearning #pytorch #computervision #edgeml #reactnative #objectdetection #yolov5
-
Intel Corporation
Unlock the full potential of your Large Language Models (LLMs) with Intel® Extension for PyTorch (IPEX) and the Intel® LLM Library for PyTorch (IPEX-LLM). Download this whitepaper to explore the optimization of LLMs, including their performance, resource utilization, and response times in real-world applications. Link - https://intel.ly/3BBJ4ey
-
Intel Corporation
Unlock the full potential of your Large Language Models (LLMs) with Intel® Extension for PyTorch (IPEX) and the Intel® LLM Library for PyTorch (IPEX-LLM). Download this whitepaper to explore the optimization of LLMs, including their performance, resource utilization, and response times in real-world applications. Link - https://intel.ly/4dwlqgQ
-
Intel Corporation
Unlock the full potential of your Large Language Models (LLMs) with Intel® Extension for PyTorch (IPEX) and the Intel® LLM Library for PyTorch (IPEX-LLM). Download this whitepaper to explore the optimization of LLMs, including their performance, resource utilization, and response times in real-world applications. Link - https://intel.ly/3BBTlHG
-
PyTorch
We are happy to announce the addition of knowledge distillation to torchtune, a PyTorch library for easily authoring, fine-tuning and experimenting with LLMs. Knowledge distillation is a technique for imparting knowledge from a larger teacher model to a smaller student model. This was a key step in Llama 3.2 pretraining. Check out how we can leverage knowledge distillation in post-training to distill Llama 3.1 8B to Llama 3.2 1B using torchtune: https://hubs.la/Q02YyqPP0
1 Comment -
PyTorch
Check out this 🔥 topic at PyTorch Conference, September 18-19 in San Francisco: "The Rise of `Transformers` in the Growing PyTorch Ecosystem" with Arthur Zucker of Hugging Face. 🤗 https://hubs.la/Q02H6By80 Explore full schedule: https://hubs.la/Q02H6CV90. Register by August 9 & SAVE US$150! https://hubs.la/Q02H6xvy0 #PyTorchConf #PyTorch #PyTorchFoundation #AI #MachineLearning #ML #DeepLearning #OpenSource #OpenSourceSoftware #OpenSourceDevelopment #OpenSourceCommunity #OSS #LinuxFoundation #events #linux #SoftwareSupplyChainNews
-
Ingonyama - Zero Knowledge Proof Hardware Acceleration
Solving Reproducibility Challenges in Deep Learning and LLMs: Our Journey Machine learning is often non-deterministic, meaning the same training or inference on identical input can yield different results. Our latest blog dives deep into how we optimized frameworks like PyTorch and llama.cpp for consistency across different hardware platforms. Why Reproducible AI Matters Our main goal is Verifiable AI, but reproducibility also helps in: - Debugging: Consistent outputs simplify bug detection and fixing. - Safety-Critical Applications: Predictability is crucial for AI in fields like medicine and automotive systems. - Performance Measurement: Reliable metrics are key for improvement. Why Achieving This is Difficult Reproducible AI is a longstanding problem. As highlighted in Duncan Riach's GTC talk, Nvidia GPUs are inherently non-deterministic when running floating point computations in parallel. For example, (A+B)+C is not necessarily equal to A+(B+C). Key Takeaways: - Hardware matters: Differences in GPU architectures can impact reproducibility, even when running the same code. - Quantization alone isn't a silver bullet: it reduces model size, but doesn’t guarantee consistent results. - Deep control Is necessary: ensuring reproducibility requires going deeper into the stack, down to kernel implementations. By paying attention to both software and hardware , we can overcome these challenges and provide reliable, deterministic outputs. We hope our insights offer a roadmap to those tackling reproducibility issues in deep learning. Read the full blogpost by Aviad Dahan here:
1 Comment -
University of San Francisco Master of Science in Data Science
🚀 Check out the latest Medium post by our student Tianyunxi Yin: "Quantization Fundamentals for Model Compression". Explore the essentials of quantization and learn how it helps in compressing machine learning models for better efficiency without compromising performance. Read more here: https://lnkd.in/gHbbR_J4 #USFMDS #DataScience #MachineLearning #ModelCompression #Quantization
-
The Linux Foundation
Lin Quiao wanted to merge PyTorch with Caffe2 in a 'zipper approach', but the APIs didn’t align. Instead, their team switched to focus entirely on #PyTorch and helped make it a high-performance, production-ready framework. Discover more at #PytorchConf 2024, running September 18-19 in San Francisco. Register now: https://hubs.la/Q02NqL230 #Pytorch #opensource #PyTorchConf
-
Barak Turovsky
𝗦𝘆𝗻𝘁𝗵𝗲𝘁𝗶𝗰 𝗗𝗮𝘁𝗮: 𝗔 𝗗𝗼𝘂𝗯𝗹𝗲-𝗘𝗱𝗴𝗲𝗱 𝗦𝘄𝗼𝗿𝗱 𝗳𝗼𝗿 𝗟𝗟𝗠 𝗧𝗿𝗮𝗶𝗻𝗶𝗻𝗴 Great article that sheds light on the limitations of using synthetic (machine-produced) data to enhance LLMs. The proliferation of synthetic data poses a challenge in differentiating between human-produced and synthetic data when crawling public sources for fresh training data. Years ago, while collecting training data for Google's AI translation models from the Internet, we encountered this exact issue: for certain smaller languages, 𝗮 𝘀𝘁𝗮𝗴𝗴𝗲𝗿𝗶𝗻𝗴 𝟱𝟬% 𝗼𝗳 𝘁𝗵𝗲 𝘁𝗿𝗮𝗻𝘀𝗹𝗮𝘁𝗲𝗱 𝗰𝗼𝗻𝘁𝗲𝗻𝘁 𝗼𝗻𝗹𝗶𝗻𝗲 𝘄𝗮𝘀 (𝗻𝗼𝗻 𝗚𝗼𝗼𝗴𝗹𝗲) 𝗺𝗮𝗰𝗵𝗶𝗻𝗲 𝘁𝗿𝗮𝗻𝘀𝗹𝗮𝘁𝗶𝗼𝗻𝘀! While leveraging synthetic data is a valuable technique for LLM training, it's crucial to understand its advantages and shortcomings and use a variety of techniques to continuously refine LLMs for your use case. It's more of an art than a science! #LargeLanguageModels #ArtificialIntelligence #AIRevolution
1 Comment -
TuringPost
QLoRA: Efficient fine-tuning of quantized LLMs It cuts memory use while fine-tuning huge LMs, using techniques like: • 4-bit NormalFloat (NF4) Quantization: Simplifies data by breaking it into chunks, like resizing images without losing quality, scaling data to fit a new range. • Double Quantization: Simplifies data storage with fewer bits, keeping accuracy. It quantizes the quantization constants, saving about 3 GB for a 65B model. • Paged Optimizers: These memory managers move data to the CPU when GPU memory is low and retrieve it when needed. This prevents crashes or slowdowns from memory spikes. Models trained on QLoRA can achieve 97.8% of ChatGPT's performance. Can something be better at reducing memory use than QLoRA? Well, it's QDoRA. Read about QDoRA and DoRA methods in our AI 101 episode: https://lnkd.in/eM7Eswgb #AI #LoRA #QLoRA #ML
-
PolyMage Labs
While we've posted in the past on PolyBlocks performance, we now report its coverage on all the 210 PyTorch models recently used in the evaluation of the PyTorch-2 paper from ASPLOS 2024 (https://lnkd.in/gj6cZCYn). These benchmarks are publicly available and drawn from HuggingFace, TIMM, and TorchBench suites. PolyBlocks can successfully compile and execute 185 of these 210 models (push button, no source changes)! These models span various AI domains with some containing nearly 25,000 tensor operators. 90% of the models that run with PyTorch's standard runtime or compiler work with PolyBlocks! Benchmarks courtesy: https://lnkd.in/gtcTNQzj
-
AIPressRoom
#Topics Improving your LLMs with RLHF on Amazon SageMaker [ad_1] Reinforcement Learning from Human Feedback (RLHF) is recognized as the industry standard technique for ensuring large language models (LLMs) produce content that is truthful, harmless, and helpful. The technique operates by training a “reward model” based on human feedback and uses this model as a reward function to optimize an agent’s policy through reinforcement learning (RL). RLHF has proven to be essential to produce LLMs such as OpenAI’s ChatGPT and Anthropic’s Claude that are aligned with human objectives. Gone are the days when you need unnatural prompt engineering to get base models, such as GPT-3, to solve your tasks. An important caveat of RLHF is that it is a complex and often unstable procedure. As a method, RLHF requires that you must first train a reward model that reflects human preferences. Then, the LLM must be fine-tuned to maximize the reward model’s estimated reward without drifting too far from the original model. In this post, we will demonstrate how to fine-tune a base model with RLHF on Amazon SageMaker. We also show you how to perform human evaluation to quantify the improvements of the resulting model. Prerequisites Before you get started, make sure yo...
-
Rabia N.
I wanted to share some insights from a recent paper from Amazon Science in which authors tackled an important challenge in language modeling: minimizing sequence truncation. In language models, truncating input sequences might lead to a loss of crucial information, especially when the models are dealing with long-context tasks such as document summarization or story generation. The authors introduce a clever solution by leveraging a heuristic from the Best Fit Bin Packing problem, an NP-hard problem, to reorder input documents and reduce truncation. The proposed approach helps retain more information compared to other strategies by optimizing how sequences are packed into the model's context window. Authors show that the bin packing solution had biggest performance gain reading comprehension (+4.7%), natural-language inference (+9.3%), context following (+16.8%), and program synthesis. They also found that best-fit packing helped prevent closed-domain hallucination. Authors claim that while the experiments primarily focused on LLM pretraining, best-fit packing is broadly applicable to fine tuning as well. https://lnkd.in/gWquw3ud
3 Comments
Explore collaborative articles
We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.
Explore MoreOthers named Benedikt S.
-
Benedikt Benediktsson
DG at SVÞ - Samtök verslunar og þjónustu (Federation of Trade & Services, Iceland)
-
Benedikt S.
--
-
Benedikt S.
--
-
benedikt s
--
-
Benedikt S
--
12 others named Benedikt S. are on LinkedIn
See others named Benedikt S.