# Core ML models **URL:** \ **Category:** Questions\ **Tags:** coreml\ **Created:** [May 30, 2021, 9:31am UTC](https://the.fmsoup.org/t/core-ml-models/2179 "2021-05-30T09:31:33Z")\ **Posts on this page:** 11\ **Page:** 1
**Author:** ![Malcolm](https://yyz2.discourse-cdn.com/flex030/user_avatar/the.fmsoup.org/malcolm/32/196_2.png) [@Malcolm](https://the.fmsoup.org/u/Malcolm)\ **Post date:** [May 30, 2021, 9:31am UTC](https://the.fmsoup.org/t/core-ml-models/2179/1 "2021-05-30T09:31:34Z")
Is anyone playing with CoreML models? I'm interested in using text analysis. In the Claris examples they have a sentiment analysis model. I want to investigate nearest neighbours, misspelled words, and so on. ---
**Author:** ![Torsten](https://yyz2.discourse-cdn.com/flex030/user_avatar/the.fmsoup.org/torsten/32/1574_2.png) [@Torsten](https://the.fmsoup.org/u/Torsten)\ **Post date:** [May 30, 2021, 10:49am UTC](https://the.fmsoup.org/t/core-ml-models/2179/2 "2021-05-30T10:49:30Z")
Not yet, but clearly a topic of interest. ---
**Author:** ![OliverBarrett](https://avatars.discourse-cdn.com/v4/letter/o/f19dbf/32.png) [@OliverBarrett](https://the.fmsoup.org/u/OliverBarrett)\ **Post date:** [May 30, 2021, 11:09am UTC](https://the.fmsoup.org/t/core-ml-models/2179/3 "2021-05-30T11:09:34Z")
I would seriously recommend you look at Python for ML. Using Pandas for data cleanup chores (a lot usually needed) and scikit-learn for ML leaves other approaches far behind. Check out the book "Hands on machine learning...."., second edition. Nearest neighbors is probably one of the very easiest (and often less useful) ML models to start with. There are also free easy-to-use sentiment analysis libraries out there though they all have some issues. One of the models, for example, can't handle misspelled words. LOL. Jump in! ---
**Author:** ![harvest](https://yyz2.discourse-cdn.com/flex030/user_avatar/the.fmsoup.org/harvest/32/231_2.png) [@harvest](https://the.fmsoup.org/u/harvest)\ **Post date:** [May 30, 2021, 12:02pm UTC](https://the.fmsoup.org/t/core-ml-models/2179/4 "2021-05-30T12:02:35Z")
really exciting topic, I lately did some research on the Apple develoment site, but all ML related items were on interpreting images which for me is of minor interest. What I'm looking for is more like clustering text by patterns or another one would be getting the OCR results from invoices put into tables... Things can be done by online services but I thought maybe the M1 architecture plus ML-libraries could bring this onsite. Maybe I am a bit early to this šŸ™‚ ---
**Author:** ![OliverBarrett](https://avatars.discourse-cdn.com/v4/letter/o/f19dbf/32.png) [@OliverBarrett](https://the.fmsoup.org/u/OliverBarrett)\ **Post date:** [May 30, 2021, 12:05pm UTC](https://the.fmsoup.org/t/core-ml-models/2179/5 "2021-05-30T12:05:20Z")
Check out the book I mentioned above on Amazon and see if what you want to do hasn't been around for quite a while... šŸ™‚ ---
**Author:** ![Malcolm](https://yyz2.discourse-cdn.com/flex030/user_avatar/the.fmsoup.org/malcolm/32/196_2.png) [@Malcolm](https://the.fmsoup.org/u/Malcolm)\ **Post date:** [May 30, 2021, 6:28pm UTC](https://the.fmsoup.org/t/core-ml-models/2179/6 "2021-05-30T18:28:31Z")
thanks for the reference @OliverBarrett. I'll dig in and have a look. I just noticed that there is a bunch of stuff on GitHub associated with the book. [GitHub - ageron/handson-ml2: A series of Jupyter notebooks that walk you through the fundamentals of Machine Learning and Deep Learning in Python using Scikit-Learn, Keras and TensorFlow 2.](https://github.com/ageron/handson-ml2) ---
**Author:** ![OliverBarrett](https://avatars.discourse-cdn.com/v4/letter/o/f19dbf/32.png) [@OliverBarrett](https://the.fmsoup.org/u/OliverBarrett)\ **Post date:** [May 30, 2021, 6:57pm UTC](https://the.fmsoup.org/t/core-ml-models/2179/7 "2021-05-30T18:57:00Z")
You are so welcome! It's really an excellent book with TONS of resources, as you found. šŸ™‚ The book suggests setting up a virtual environment for each project so you keep libraries and such separate. Regarding "Pandas"...this is such a powerful library (I use it for linear algebra - ML related stuff often), I can also recommend "Python for Data Analysis". This book has so many excellent examples. Drop a note and let me know how it's going. Enjoy!!!!! ---
**Author:** ![FMUSER01](https://avatars.discourse-cdn.com/v4/letter/f/a8b319/32.png) [@FMUSER01](https://the.fmsoup.org/u/FMUSER01)\ **Post date:** [August 13, 2021, 3:36pm UTC](https://the.fmsoup.org/t/core-ml-models/2179/8 "2021-08-13T15:36:35Z")
Machine Learning and ā€œApple, you have some ā€™splaining to do!ā€ A hard-hitting but insightful video interview by Joanna Stern from The Wall Street Journal of Appleā€™s software chief Craig Federighi regarding Appleā€™s iPhone child protection features and methods. [Link to WSJ video interview](https://www.wsj.com/video/series/joanna-stern-personal-technology/apples-software-chief-explains-misunderstood-iphone-child-protection-features-exclusive/573D76B3-5ACF-4C87-ACE1-E99CECEFA82C?mod=hp_lead_pos5) Federighi tries, Stern actually delivers. ---
**Author:** ![tonni6030](https://avatars.discourse-cdn.com/v4/letter/t/b19c9b/32.png) [@tonni6030](https://the.fmsoup.org/u/tonni6030)\ **Post date:** [July 10, 2025, 12:40pm UTC](https://the.fmsoup.org/t/core-ml-models/2179/9 "2025-07-10T12:40:37Z")
Yes, Iā€™ve been trying out CoreML with FileMaker recently. The sentiment analysis model Claris shared is a good starting point, but if you're looking into things like nearest neighbours or handling misspelled words, you might want to look into custom NLP models or even use something like a word embedding model. Iā€™ve found that integrating these models through a small script or external function works better for more advanced text processing in [software development](https://applicontech.com/) projects. It takes a bit of setup, but it opens up a lot more flexibility than just using the built-in models. Let me know if you want help setting any of that up. ---
**Author:** ![Malcolm](https://yyz2.discourse-cdn.com/flex030/user_avatar/the.fmsoup.org/malcolm/32/196_2.png) [@Malcolm](https://the.fmsoup.org/u/Malcolm)\ **Post date:** [July 10, 2025, 9:01pm UTC](https://the.fmsoup.org/t/core-ml-models/2179/10 "2025-07-10T21:01:25Z")
I'd be interested to see what setup is needed for that. I'm sure others would be too. The Soup is highly searched via Google. If you posted in the Resources channel it could be a how-to. ---
**Author:** ![OliverBarrett](https://avatars.discourse-cdn.com/v4/letter/o/f19dbf/32.png) [@OliverBarrett](https://the.fmsoup.org/u/OliverBarrett)\ **Post date:** [July 12, 2025, 12:36pm UTC](https://the.fmsoup.org/t/core-ml-models/2179/11 "2025-07-12T12:36:06Z")
I just use an OLLAMA totally LOCAL model. It's free and easily accessible by FileMaker, as long as you have FileMaker 16+. šŸ˜ƒ You can also access other AI models easily using INSERT FROM URL with older versions of FileMaker. (No AI Hype Train Here...)