Hi all. A small question regarding encoding the position of inputs to a transformer model.
How would you encode a set of sequences to a (bidirectional) transformer? For a sequence we have positional encodings. For a set we can just work without them. What about a set of sequences {s\_1, …, s\_n}, where each s\_1, …, s\_n is a sequence, but their relative order does not matter?
One of the gaps I kept hitting in SEO audits was the **lack of a clean visual archive** of SERPs across time — especially for high-value commercial queries.
So I built a desktop tool (macOS, Electron + Chromium) that:
\- Accepts a keyword list
\- Automates browser sessions
\- Captures full-page SERP screenshots
\- Saves them locally for comparison
It’s not scraping — it captures what a real user sees, including any A/B variations or local result shifts. Helpful when analyzing SERP volatility or preparing reports for non-technical clients.
Curious how others are handling **visual SERP tracking** — and whether there’s a better way to structure this process.
Can share a demo or the tool if anyone’s interested.
Building a custom options screener doesn’t have to be a daunting task. By choosing the right API and starting simple, you create a tool that can greatly enhance your or your family’s trading strategies. Remember, it’s all about making trading information accessible and tailored to personal needs.
So go ahead, take that first step. You might just find that building tools for trading is not only useful but also quite rewarding. Happy coding!
Hello everyone! Just wanted to share a new SEO update because this community has helped me so much. I posted here a few months back about taking SEO more seriously for my Boston-based video production company, and the response was honestly so kind that I wanted to post again with updated results and more gratitude. I also wanted to just circle back to shed some light on things that have been working really well for me in the last three months!
I’m now about 16 months into the real SEO grind – consistent blogging, optimizing titles and H1s, internal linking, structured data – all that good stuff.
Before I jump in, I do want to mention that I have some photos from Google Search Console below so that you can see the journey. Also, I know it’s not all about impressions – we’ve been booking a lot of gigs via google search and this all has lead to more clicks!
So back in March(ish) when I posted I mentioned that I was doing the following:
* Writing blog posts that got attention on google
* Building individual service pages for everything we do
* Using Google Search Console and GA4 to figure out what was working
* Fixing technical stuff in Squarespace
* Learning how to write and inject structured data. To be honest, I use chatGPT to help me write this code. I then take it to Google Rich Results and test it and go back and forth with ChatGPT to make sure it’s perfect.
* Internally linking like a maniac
* Getting every client to leave a Google review
* Asking other video production companies around the country to take meetings with me and learn more about who they are. If we think we’re a good partner to work together on something in the future we will both write a really detailed blog about one another for backlinks. I’m very careful about this and admit that backlinks are my weakest subject
* Updating our Google Business Profile weekly with posts and photos
Since then I have seen a lot of growth, specifically around May. This huge jump has made me wonder what I’m doing differently. I think I can credit the recent growth to the following:
* A **few blogs that have really hit it off** \- these are gear related blogs that have a lot of traffic. Are they potential clients? no but I think it is helping with authority!
* I have been **implementing schema code** on all our pages and every single blog I write!
* **consistency**
* **I changed my home page title** from “Boston Based Video Production | Bunker Hill Media” to “Bunker Hill Media: Video Production That Drives Proven Engagement” and I think this is helping with clicks. IMO, I love that title because it’s true!
I feel like I’m learning something new every day, especially when it comes to backlinks (still my weakest area). But this community has made SEO feel way less intimidating and way more collaborative!
I think the one thing I’m scared of a bit now is that I’m noticing I’m dipping in rankings. I seem to be staying right around the same place for what I want to be there for “video production in Boston” – right around 4-7 on google (up from 90th last year!). But working on keeping that up and improving CTR!
Thanks again for all the support xoxoxo!
In conclusion, while anyone’s travel choice is deeply personal and conditional on comfort levels, Monastir offers more than fear might suggest upon first hearing mixed reviews. With sensible precautions and an open heart, your family vacation here could be both enriching and safe, filled with remarkable stories to bring back home.
Hi there!
I am planning a semi last minute solo trip to Italy the last week of August/ beginning of September. There are so many incredible cities, but have come narrowed it down to the following stops:
Fly into Rome, then to Florence, then to Cinque Terre, then would love to Bergamo/Dolomites area. Wine, good food, and great views are my goal.
1. Public transportation seems to be great over there, but with this itinerary should I rent a car? The only spot that concerns me for parking, etc. mainly in Cinque Terre is Cinque Terre since I don’t plan to spend too much time in the larger cities
2. Any recommendations on spots that I am missing? Love the walkability, hiking aspect of many of the locations.
3. Any recommendations on where to stay in Cinque Terre or Dolomites area? Any amazing experiences at hotels? Preferences between hotels vs air bnbs?
4. Has anyone used any app, program, watched any youtube videos that helped them learn words and phrases to help get around over there?
Any and all advice/ recommendations are welcome! First time to Europe so incredibly excited. Thanks!!
Solana’s block capacity jumped 20% to 60M compute units since July 21, 2025, allowing more transactions per block. This move, backed by Solana docs, tackles past congestion (e.g., 2024 meme coin spike) to cut fees and boost UX
I am pleased to introduce [`treemind`](https://github.com/sametcopur/treemind/), a high-performance Python library for interpreting tree-based models.
Whether you’re auditing models, debugging feature behavior, or exploring feature interactions, `treemind` provides a robust and scalable solution with meaningful visual explanations.
* **Feature Analysis** Understand how individual features influence model predictions across different split intervals.
* **Interaction Detection** Automatically detect and rank pairwise or higher-order feature interactions.
* **Model Support** Works seamlessly with LightGBM, XGBoost, CatBoost, scikit-learn, and perpetual.
* **Performance Optimized** Fast even on deep and wide ensembles via Cython-backed internals.
* **Visualizations** Includes a plotting module for interaction maps, importance heatmaps, feature influence charts, and more.
**Installation**
pip install treemind
**One-Dimensional Feature Explanation**
Each row in the table shows how the model behaves within a specific range of the selected feature.
The `value` column represents the average prediction in that interval, making it easier to identify which value ranges influence the model most.
| worst_texture_lb | worst_texture_ub | value | std | count |
|——————|——————|———–|———-|———|
| -inf | 18.460 | 3.185128 | 8.479232 | 402.24 |
| 18.460 | 19.300 | 3.160656 | 8.519873 | 402.39 |
| 19.300 | 19.415 | 3.119814 | 8.489262 | 401.85 |
| 19.415 | 20.225 | 3.101601 | 8.490439 | 402.55 |
| 20.225 | 20.360 | 2.772929 | 8.711773 | 433.16 |
**Feature Plot**
#
**Two Dimensional Interaction Plot**
The plot shows how the model’s prediction varies across value combinations of two features. It highlights regions where their joint influence is strongest, revealing important interactions.
# Learn More
* Documentation: [https://treemind.readthedocs.io](https://treemind.readthedocs.io)
* Github: [https://github.com/sametcopur/treemind/](https://github.com/sametcopur/treemind/)
* Algorithm Details: [How It Works](https://treemind.readthedocs.io/en/latest/algorithm.html)
* Benchmarks: [Performance Evaluation](https://treemind.readthedocs.io/en/latest/experiments/experiment_main.html)
Feedback and contributions are welcome. If you’re working on model interpretability, we’d love to hear your thoughts.
Ash is an AI therapy trained on actual therapy data. Interesting!
