
Founder @MantisNLP 🪲 Data scientist with a focus on Open source 💻 Reproducible 🔬 and Ethical ⚖️ code. Member of @PyData and @DataKind
London
Joined May 2009
- Tweets1.2K
- Following310
- Followers342
- Likes429
Pinned Tweet
✨ Here is a project I have been working on for the last 2 years with the @wellcometrust
We trained a transformers 🤖 based model to tag biomedical grants with MeSH tags (29K labels) and shared in the huggingface hub🤗
Model is trained for 4 days using 8 GPUs and scores 63% f1
Nick Sorros retweeted
Athens on the list of European cities with the highest per-capita density of engineers for a number of specialisations, according to Sequoia’s Atlas report
AI, data science, front-end, security, systems, etc
atlas.sequoiacap.com
✍️ New blog post out about using SageMaker for inference. Hope you like it 🙏🏻
This Post is from an account that no longer exists. Learn more
1
100
I think LLMs will play an important role for accelerating prototyping NLP solutions but for many use cases, smaller / optimised models will be a much better fit in the near future. Thoughts?
This Post is from an account that no longer exists. Learn more
1
113
Here is a nice CLI tool we created on top of the @huggingface inference endpoints which allow you to deploy any transformer model from the command line or a CI/CD pipeline.
Kudos mainly to @m_a_upson 👏
This Post is from an account that no longer exists. Learn more
🔥 Do we even need to annotate anymore in the prototype phase?
✍️ Just published a new blog about prototyping with Large Language Models, let me know what you think.
This Post is from an account that no longer exists. Learn more
87
📚 Large Language Models will definitely change how we all learn. @duolingo and @khanacademy offer a glimpse of what's to come but I am sure this is just the tip of the iceberg 🧊
This Post is from an account that no longer exists. Learn more
1
97
I am struggling to find a comparison of bloomz with other instruction models like flan and gpt3/4? Any tips from @huggingface or friends is more than welcome 🙏🏻
1
88
As it turns you can use an LLM to teach another LLM how to follow instructions 🤯
What's even better, the cost to do that is under 1000$ 🔥
This Post is from an account that no longer exists. Learn more
2
114
✨Large Language Models like #GPT4 can speed up your data creation process in applications where you need to build smaller and more predictable and performant models 🚀
This Post is from an account that no longer exists. Learn more
1
81
✨ GovGPT is here. We created a #ChatGPT like demo 🕹 on top of gov.uk data. We have actively worked on reducing hallucinations as much as possible.
Read our blog below or try it here chat.mantisnlp.com/
This Post is from an account that no longer exists. Learn more
1
71
If you have seen flamingo 🦩 from @deepmind, this is an open source equivalent from @salesforce available in the @huggingface that's worth checking out ✨
This Post is from an account that no longer exists. Learn more
1
86
💥 New blog out, cheekily named "8 examples is all you need" about SetFit from @huggingface 🤗
When working in a text classification problem, its definitely keeping this technique in mind 👌
This Post is from an account that no longer exists. Learn more
102
Pairing Large Language Models with external tools open ups many more applications than what initially comes to mind when interacting with GPT3 or #chatgpt
This Post is from an account that no longer exists. Learn more
87
If you want to use Large Language Models like GPT3 to annotate data so that you can train a more performant 🚀 and smaller 🔥 model, definitely check these new recipes from @explosion_ai
This Post is from an account that no longer exists. Learn more
We recently switched to using @huggingface inference endpoints as it makes it much easier to deploy and manage the models for our clients. We also created a nice CLI tool called hugie 🐻 to make inference endpoints available from the command line 🚀
This Post is from an account that no longer exists. Learn more


