Very cool discovery, thank you! I like the flexibility of the 0-shot approach. Not to mention almost all trained NER models online for CV analysis all seem to use the same job posting data set from 2018.
Now we only need to get the relationships between the entities :) great video Mark! Did you try it with bigger text sizes like a larger wikipedia article or a book?
Now I wonder where we could store those entities :D I think it kinda maxes out at around 2,000 characters. But I suppose if you would chop it up into chunks and run it again it's gonna work?
is there something like this for text classification? I have a pdf and a list of topics such as [profit, environment, competitors…]. I need to classify each page in my 500 paged pdf file to know which topics been discussed in each page
It might be worth looking at the LLMware models - github.com/llmware-ai/llmware/tree/main/examples/SLIM-Agents I did a video about a few of them here - ruclips.net/video/Lb4_z-UI80A/видео.html
Hey, yep. I'm using this - pypi.org/project/ipython/ And then for colouring when I print stuff I'm using this - rich.readthedocs.io/en/stable/console.html
Very cool discovery, thank you! I like the flexibility of the 0-shot approach. Not to mention almost all trained NER models online for CV analysis all seem to use the same job posting data set from 2018.
Thank you so much
Now we only need to get the relationships between the entities :) great video Mark! Did you try it with bigger text sizes like a larger wikipedia article or a book?
Now I wonder where we could store those entities :D
I think it kinda maxes out at around 2,000 characters. But I suppose if you would chop it up into chunks and run it again it's gonna work?
Great video, thanks! could you please help me with loading the model from a local path?
using path with local_files_only=True failed for me.
is there something like this for text classification? I have a pdf and a list of topics such as [profit, environment, competitors…].
I need to classify each page in my 500 paged pdf file to know which topics been discussed in each page
It might be worth looking at the LLMware models - github.com/llmware-ai/llmware/tree/main/examples/SLIM-Agents
I did a video about a few of them here - ruclips.net/video/Lb4_z-UI80A/видео.html
Hey Mark, as always, short and powerful demo : how does it handle i8n ? I would like to play with some french data ;-p
Good question. I haven't tried that yet!
There’s a multilingual version as well
Has anyone use this via a cloud endpoint? Its so custom that is not even compatible with hugging face cloud endpoint
I haven't tried it yet, but we could deploy it to something like Beam.Cloud? www.beam.cloud/
Can you explain what you use to run jupyter on the terminal?
Hey, yep. I'm using this - pypi.org/project/ipython/
And then for colouring when I print stuff I'm using this - rich.readthedocs.io/en/stable/console.html
@@learndatawithmark I guess that's worth another python video :)