Machine Learning Engineer - @HuggingFace / Paris 🇫🇷
Paris
Joined May 2012
- Tweets316
- Following313
- Followers1.6K
- Likes829
Morgan Funtowicz retweeted
🤗 Transformers has been built by, with, and for the community.
Reaching 100k ⭐ on GitHub is a testament to ML's reach and the community's will to innovate and contribute.
To celebrate, we highlight 100 incredible projects in transformers' vicinity.
github.com/huggingface/trans…
ALT Hugging Faces celebrating the 100k milestone. Highlights that there are many Hugging Faces, as reaching this milestone is really thanks to a huge community building with this tool.
Morgan Funtowicz retweeted
ONNX Runtime inference for Seq2Seq models has never been so easy ! 🚀
With 🤗 Optimum, you can now export your T5 model to the ONNX format and perform ONNX Runtime inference using 🤗 Transformers pipelines.
➡️ Start here huggingface.co/docs/optimum/…
Morgan Funtowicz retweeted
🗣 I'll be speaking tomorrow Friday 1pm about 🤗 transformers acceleration with Optimum and @onnxruntime - join us in Mountain View or via the livestream!
😍😍ONLY 1 MORE DAY TILL THE ONNX COMMUNITY DAY
Have you registered? Click here 👇
aka.ms/onnxcommunityday-june…
@onnxai @Oracle @awscloud @huggingface @Adobe @IBMData @AMD @jeffboudier @kalra05nikhil @Ale_lrg07 #opensource #MLOps
Morgan Funtowicz retweeted
Recently, @huggingface released a tool called Optimum. It integrates @onnxruntime and makes it easy to optimize training and inference. Come hang out, ask questions, and learn with the engineers from both teams who build awesome open-source tools:
msft.it/6018bc5cE
Morgan Funtowicz retweeted
Are GPUs the only future for fast, efficient Transformers workload? 🤔
No, not only!
We created a blog on how to set up a Deep Learning environment for @Hugging Face Transformers with @HabanaLabs Gaudi on @awscloud 🏎💨💨
📕philschmid.de/getting-starte…
⭐️ github.com/philschmid/deep-l…
ALT code snippet
Morgan Funtowicz retweeted
Important news! @huggingface is now compliant with SOC 2 Type 1. Our customers (especially in regulated industries) expect a high level of security and privacy for their data and models, and this is an important first step! #MachineLearning #Security huggingface.co/
Morgan Funtowicz retweeted
You can now accelerate inference by applying quantization to models from the @huggingface Hub 🔥
➡️ With 🤗 Optimum, you can easily apply static and dynamic quantization on your model before exporting it to the ONNX format 🤯
Start here 👉 huggingface.co/docs/optimum/…
Morgan Funtowicz retweeted
Hugging Face Infinity is a containerized solution for customers to deploy end-to-end optimized inference pipelines on any infrastructure🧱
We created a case study focused on modern CPUs including a benchmark with 192 different experiments 🧪
huggingface.co/blog/infinity…
Morgan Funtowicz retweeted
Lately I've been exporting 🤗 Transformers models to ONNX and the API that @MorganFunto designed is really nice!
Here's an example to export DistilBERT to ONNX and then run inference with @onnxruntime -- as you can see it's just one line of code to export a supported model 🤯
Morgan Funtowicz retweeted
😍🤗 – at Miami, FL
We're ready for you!
Join me and team @huggingface at the #AISummit NYC Wed 12/8 Thu 12/9 and ask us anything! 👋🗽
I will speak about accelerating Transformers to millisecond latency and our work with @IntelAI
⏰ Wed 1:45pm EST
newyork.theaisummit.com/
Morgan Funtowicz retweeted
On Friday @huggingface officially achieved the #AWS Machine Learning competency. This is a very tough competency to get with a limited amount of partners having the qualifications to achieve it! 🔥 🎉
Morgan Funtowicz retweeted
Infinity by @huggingface lets you optimize the prediction latency of #Transformers up to 10x... with a single #Docker command! Check out my end to end demo at youtu.be/ngB8VBbVRd4 #MachineLearning #MLOps
Morgan Funtowicz retweeted
New @huggingface blog post! Am I excited by mega-models? No. Here's why, and what I think we should work on instead. huggingface.co/blog/large-la… #MachineLearning #DeepLearning #NLP
Morgan Funtowicz retweeted
⏳ We launch 🤗 Infinity, our new inference solution, in 4️⃣ days - register for the event today!
Watch how @otto_tech sped up inference 10X for semantic search with 🤗 Infinity - and set it up in minutes!🚀
📅 Tue 9/28 - 9am PT | 12pm ET | 6pm CET
app.livestorm.co/hugging-fac…
Morgan Funtowicz retweeted
You can now use TensorBoard on private Hugging Face repos. There are over 500 models using TensorBoard on @huggingface, and many dozens of private repos now, too!
🤗 Take a look 👇
Did you know that if you push *any* TensorBoard traces to @huggingface, you get an automatically deployed, free TensorBoard instance that helps you during model development?
Now also for private repos!
Morgan Funtowicz retweeted
We are proud to collaborate with @IntelAI to Accelerate AI for Production! 🤝🦾
Check out how to easily quantize and prune models with our new 🤗 Optimum library, integrating Intel LPOT:
huggingface.co/hardware
Reply if you spotted this billboard on US 101 by SFO this week! 😎
Morgan Funtowicz retweeted
We are excited to team up with @graphcoreai, to make training of 🤗 Transformers on cutting-edge IPUs super easy!
With our new library 🤗 Optimum, using IPUs will become plug-and-play.
Read all about it in this guest blog post from the Graphcore team:
huggingface.co/blog/graphcor…
Morgan Funtowicz retweeted
📢 Introducing 🤗 Optimum
A new open source library to optimize 🤗Transformers for production performance. 🏎
Quantize, Prune, Optimize models easily, targeting hardware from our partners @intel @graphcoreai @Qualcomm! 🤩
huggingface.co/blog/hardware…
Morgan Funtowicz retweeted
📢 Improved @onnxruntime support in 🤗 Transformers!
Exporting any of the 60+ Transformers architectures from @Pytorch to @onnxai is now supported with customizable configs! 🔥
ONNX is now in its own package:
💫 transformers.onnx 💫
Check it out here: huggingface.co/transformers/…
Morgan Funtowicz retweeted
With #TensorRT 8, we achieved 1ms inference latency on BERT, and we’re building this performance into a new service later this year!
This, and more from @jeffboudier, in the @NVIDIAAI press release nvda.ws/3rjuNu1
Today NVIDIA launched #TensorRT 8 with BERT-Large inference down to 1 ms, enabling developers to build the world’s best-performing search engines, ad recommendations and chatbots and offer them from the cloud to the edge. #NLP nvda.ws/3rjuNu1