←
PyDataMTL74 posts
PyDataMTL avatar

PyDataMTL

@PyDataMTL

We ❤️ Data Science, Machine Learning, Deep Learning and meet once a month to learn from each other. Brought to you by @mariakhalusova and @alex000kim

◷ Joined January 2019▧ 8 Photos and videos
PyDataMTL avatar
PyDataMTL@PyDataMTL·24 Feb 2021

Next, I will show how information extraction from social media data can be looked through the lens of an abstraction called digital social trace data (DSTD) which allows us to utilize the graph structure of the data for developing new information extraction tasks. (3/5)

PyDataMTL avatar
PyDataMTL@PyDataMTL·24 Feb 2021

sarcasm, or abuse) from social media text using semi-supervised, multi-task, and active learning to more accurately identify phrases, named entities, and classification. I will introduce an open source tool called SocialMediaIE. (2/5)

PyDataMTL avatar
PyDataMTL@PyDataMTL·24 Feb 2021

Finally, tomorrow we will learn from @TheShubhanshu. His talk is called "Understanding digital social trace data via Information Extraction" Abstract: I will describe how to efficiently extract information (e.g. phrases, named entities, and categories like topics, sentiment,(1/5)

PyDataMTL avatar
PyDataMTL@PyDataMTL·23 Feb 2021

The second talk on the agenda of our upcoming NLP meetup is "Learning to translate with JoeyNMT" by @KreutzerJulia. Abstract: Machine translation is one of the most impactful NLP applications. Despite many toolkits for it being open-sourced, 1/3

PyDataMTL avatar
PyDataMTL@PyDataMTL·22 Feb 2021

and show how to apply them on an example of fine-tuning. In particular, we will see how the variants of ZeRO are currently implemented in the Transformers library, by the use of fairscale and deepspeed. 4/4

PyDataMTL avatar
PyDataMTL@PyDataMTL·22 Feb 2021

In this talk, we will dive into the techniques one can use to reduce the GPU-memory used when fine-tuning a model (gradient accumulation, gradient checkpointing, model parallelism, and the more recent ZeRO/ZeRO-offload developments), 3/4

PyDataMTL avatar
PyDataMTL@PyDataMTL·22 Feb 2021

This, in turn, has led researchers to develop larger and larger language models, which give better results than their smaller versions but can be hard to use without powerful (and expensive) hardware. 2/4

PyDataMTL avatar
PyDataMTL@PyDataMTL·22 Feb 2021

The first talk of our next meetup: "Fine-tuning a large language model without your own supercomputer" by @GuggerSylvain Abstract: State-of-the-art results in NLP involve more and more the use of a generic pretrained language model fine-tuned on a specific downstream task. 1/4