TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

Show HN: Structured Outputs with Deepseek R1

7 pointsby aaronvg4 months ago
This is our interactive playground that uses our framework (BAML <a href="https:&#x2F;&#x2F;github.com&#x2F;BoundaryML&#x2F;baml">https:&#x2F;&#x2F;github.com&#x2F;BoundaryML&#x2F;baml</a>) to do structured outputs with R1, without the use of tool-calling APIs or fine-tuning required.<p>We have a more fleshed out playground at <a href="https:&#x2F;&#x2F;promptfiddle.com" rel="nofollow">https:&#x2F;&#x2F;promptfiddle.com</a> as well.<p>BAML is a DSL for prompts, where prompts are modeled as functions. Our compiler transforms your LLM function declaration into the relevant API call, and will also parse the output for you.<p>We serialize using &quot;type definitions&quot; instead of using json schemas, since they are more efficient and easier to understand for models. We talk more about why here: <a href="https:&#x2F;&#x2F;www.boundaryml.com&#x2F;blog&#x2F;type-definition-prompting-baml">https:&#x2F;&#x2F;www.boundaryml.com&#x2F;blog&#x2F;type-definition-prompting-ba...</a><p>We ran the Berkeley Function Calling Benchmark some months ago on our approach and achieved really good results <a href="https:&#x2F;&#x2F;www.boundaryml.com&#x2F;blog&#x2F;sota-function-calling?q=0">https:&#x2F;&#x2F;www.boundaryml.com&#x2F;blog&#x2F;sota-function-calling?q=0</a>

1 comment

huevosabio4 months ago
Is there a reason for why you wouldn&#x27;t use something like Outlines or Xgrammar to force a grammar on the output?
评论 #42846983 未加载