Hacker News: parthsareen

New comment by parthsareen in "Qwen 3.8 27B"

parthsareen — Fri, 14 Aug 2026 21:37:37 +0000

Hi! From Ollama here - you can run: ollama run qwen3.8 (or if on mac qwen3.8:27b-mlx)

New comment by parthsareen in "Claude Code: connect to a local model when your quota runs out"

parthsareen — Thu, 05 Feb 2026 01:39:05 +0000

Also recently added ollama launch claude if you want to connect to cloud models from there :)

New comment by parthsareen in "Ask HN: A good Model to choose in Ollama to run on Claude Code"

parthsareen — Tue, 27 Jan 2026 00:49:11 +0000

Hey! One of the maintainers of Ollama. 8GB of VRAM is a bit tight for coding agents since their prompts are quite large. You could try playing with qwen3 and at least 16k context length to see how it works.

Building Reliable AI Agents

parthsareen — Fri, 09 Jan 2026 20:22:59 +0000

Article URL: https://parthsareen.com/writings/building-reliable-ai-agents/

Comments URL: https://news.ycombinator.com/item?id=46558809

Points: 1

# Comments: 0

New comment by parthsareen in "A guide to local coding models"

parthsareen — Mon, 22 Dec 2025 05:14:50 +0000

How much ram are you running with? Qwen3 and gpt-oss:20b punch a good bit above their weight. Personally use it for small agents.

New comment by parthsareen in "A guide to local coding models"

parthsareen — Mon, 22 Dec 2025 05:10:40 +0000

You're welcome to go through the source: https://github.com/ollama/ollama/

New comment by parthsareen in "A guide to local coding models"

parthsareen — Mon, 22 Dec 2025 05:09:30 +0000

Desktop app is open-source now.

How to Train an LLM: Part 1

parthsareen — Tue, 11 Nov 2025 19:38:59 +0000

Article URL: https://omkaark.com/posts/llm-1b-1.html

Comments URL: https://news.ycombinator.com/item?id=45891801

Points: 20

# Comments: 3

New comment by parthsareen in "Ollama Web Search"

parthsareen — Thu, 25 Sep 2025 20:58:28 +0000

Since we shipped web search with gpt-oss in the Ollama app I've personally been using that a lot more especially for research heavy tasks that I can shoot off. Plus with a 5090 or the new macs it's super fast.

New comment by parthsareen in "Ollama Web Search"

parthsareen — Thu, 25 Sep 2025 20:44:09 +0000

Hi - author of the post. Yes it does! The "build a search agent" example can be used with a local model. I'd recommend trying qwen3 or gpt-oss

New comment by parthsareen in "Ollama Web Search"

parthsareen — Thu, 25 Sep 2025 20:43:22 +0000

Hey! Author of the blogpost and I also work on Ollama's tool calling. There has been a big push on tool calling over the last year to improve the parsing. What's the issues you're running into with local tool use? What models are you using?

New comment by parthsareen in "Sampling and structured outputs in LLMs"

parthsareen — Thu, 25 Sep 2025 06:23:30 +0000

That's a great idea. Going to try this next :)

New comment by parthsareen in "Sampling and structured outputs in LLMs"

parthsareen — Thu, 25 Sep 2025 06:22:49 +0000

Hey! I'm the author of the post. We haven't optimized sampling yet so it's running linearly on the CPU. A lot of SOTA work either does this while the model is running the forward pass or does the masking on the GPU.

The greedy accept is so that the mask doesn't need to be computed. Planning to make this more efficient from either ends.

New comment by parthsareen in "Sampling and structured outputs in LLMs"

parthsareen — Tue, 23 Sep 2025 19:01:58 +0000

Thank you! Maybe not "perfect" but near-perfect is something we can expect. Models like the Osmosis structure which just structure data inspired some of that thinking (https://ollama.com/Osmosis/Osmosis-Structure-0.6B). Historically, JSON generation has been a latent capability of a model rather than a trained one, but that seems to be changing. gpt-oss was particularly trained for this type of behavior and so the token probabilities are heavily skewed to conform to JSON. Will be interesting to see the next batch of models!

New comment by parthsareen in "Sampling and structured outputs in LLMs"

parthsareen — Tue, 23 Sep 2025 18:01:02 +0000

Thanks for posting! Didn't expect this to get picked up – it was a bit of a draft haha. Happy to answer questions around structured outputs :)

New comment by parthsareen in "Structured Outputs with Ollama"

parthsareen — Mon, 09 Dec 2024 22:24:29 +0000

Yes! I have checked guidance out, as well as a few others. Planning to refactor sampling in the near future which would include improving using grammars for sampling as well. Thanks for sharing!

New comment by parthsareen in "Structured Outputs with Ollama"

parthsareen — Sat, 07 Dec 2024 17:57:57 +0000

The constraints will always be met. It’s the data inside that might be inaccurate. YMMV with smaller models in that sense.

New comment by parthsareen in "Structured Outputs with Ollama"

parthsareen — Sat, 07 Dec 2024 17:31:04 +0000

Hey! Author of the blog here. The current implementation uses llama.cpp GBNF which has allowed for a quick implementation. The biggest value-add at this time was getting the feature out.

With the newer research - outlines/xgrammar coming out, I hope to be able to update the sampling to support more formats, increase accuracy, and improve performance.

New comment by parthsareen in "Structured Outputs with Ollama"

parthsareen — Sat, 07 Dec 2024 17:26:39 +0000

Hey! Author of the post and one of the maintainers here. I agree - we (maintainers) got to this late and in general want to encourage more contributions.

Hoping to be more on top of community PRs and get them merged in the coming year.

New comment by parthsareen in "Structured Outputs with Ollama"

parthsareen — Sat, 07 Dec 2024 04:02:27 +0000

This looks really useful. Thank you!