Points: 2

# Comments: 2

This does not proxy via another service that might go down or be beholden to whims of some board.

We wrote logos-shift for our own needs and are open sourcing it.

What it does:

- Save input/output data, no proxy, no slowdowns, double buffering, separate thread

- Finetune your own model, or we can for a fee.

- Auto rollout your model when ready.

We recommend keeping your LLM at 80/20 for most cases and rollout to 100 on failure.

How to use it:

pip install logos_shift_client

from logos_shift_client import LogosShift

# Initialize with your API key (without if you just want the local copy)

logos_shift = LogosShift(api_key="YOUR_API_KEY")

# Instrument your function

@logos_shift()

def add(x, y):

    return x + y

# Optionally, provide feedback logos_shift.provide_feedback(result['bohita_logos_shift_id_123', "success")

We can manage all this for you of course, but its not hard to do it yourself.

Save data on our infra for free for upto 1GB.

$10 per month for after.

Download your data anytime. No lock in.

Feel free to ask questions and how can we make this easier for you.

Comments URL: https://news.ycombinator.com/item?id=38364279

Points: 1

# Comments: 0

Last I looked, the industry was trying to roll out some protocols to block open source efforts.

Is there a plan to tackle that?

However, once you know what you want, you should finetune and switch to your finetuned LLM.

Logos Shift Cloud handles this automatically

All you need is to add this to your call:

``` @logos_shift(dataset="story") def get_story(model, messages): """Generates a story""" completion = expensiveLLM.create(model=model, messages=messages) story_d = {'story': completion['choices'][0]['message']} return story_d ```

This will automatically log all inputs and outputs to this function unobtrusively. Once your model is ready, it will automatically A/B rollout your finetuned LLM.

No code changes necessary. Enjoy faster and cheaper results.

No proxies so you don't add one more dependency. The client code is open sourced at https://github.com/virevolai/logos-shift-client/ which includes the instrumentation, router for switching, and exactly how this works.

It's only 530 lines of code. No complex libraries.

If you do not want to use the Cloud solution, you can always just use Logos Shift for saving to local host and finetune yourself.

``` logos_shift = LogosShift(api_key=None, filename="api_calls.log") ```

Let us know how we can make this experience better.

Comments URL: https://news.ycombinator.com/item?id=38070726

Points: 1

# Comments: 0

General purpose LLMs are very expensive. You can finetune and rollout your own at a fraction of the cost. We can do it for you too.

Highlights of Logos Shift:

Automated Rollout: With just a single line of code, you can integrate your fine-tuned LLMs, and we handle the rest. Our automated rollout system ensures that your model is deployed smoothly, and we'll notify you the moment it's production-ready.

Automated A/B Testing: Test the waters with our effortless A/B rollout, activated as soon as your model achieves optimal readiness. Skip the manual steps and dive straight into performance evaluation. This code is in the open source, no black boxes.

Feedback-Driven Fine-Tuning: Continuously refine your model's performance based on real outcomes. Your feedback directly influences future rollouts and fine-tuning, aligning model performance with your actual needs.

For Hackers and Developers:

Build Your Own or Leverage Bohita: The power of choice is in your hands. Save your data locally and finetune your model on your terms, or let Bohita’s servers amplify your capabilities with our full suite of services.

Your Keys, Your Control: Your OpenAI key remains securely in your possession, ensuring complete confidentiality and integrity of your data.

Open Source and Community-Driven: Logos Shift is not just a tool; it’s a community. Dive into the open-source code, tailor it to your needs, and join a network of hackers and developers passionate about pushing the boundaries of LLM integration.

Getting Started is a Breeze: Installation is as simple as running pip install logos_shift_client. Dive into our documentation and explore the GitHub repository to kickstart your journey.

We can't wait to see what you'll build with Logos Shift. Questions, feedback, or high-fives? We’re all ears!

Happy hacking, The Bohita Team

https://api.bohita.com

Comments URL: https://news.ycombinator.com/item?id=38027295

Points: 3

# Comments: 0

But credit card receipts show 18% off the price + tax, effectively 20-25%

Most don’t do the math, just pay.

Seems like some X can now run in wasm should come with disclaimer (includes Linux)