Hugging Face's logo

Join the conversation

Join the community of Machine Learners and AI enthusiasts.

Sign Up
Back to feed
sequelbox 
posted an update 15 days ago
Post
2179
NEW RELEASE: Shining Valiant 3 now available for openai/gpt-oss-20b!

- Cutting edge science-reasoning: sequelbox/Celestia3-DeepSeek-R1-0528 for physics, biology, chemistry, compsci, astronomy, Earth science, and information theory.
- AI to build AI: the all-new sequelbox/Mitakihara-DeepSeek-R1-0528 dataset for high-quality reasoning performance on AI, MLOps, math and CUDA, complex adaptive and agentic systems, cognition, logic, linguistics, simulation, knowledge management, and more!
- Creative reasoning and general chat performance supplemented with sequelbox/Raiden-DeepSeek-R1

Get the new SV3: ValiantLabs/gpt-oss-20b-ShiningValiant3

This is our first release for the new openai/gpt-oss-20b - we're hoping to support this model with more releases going forward.

We're also excited to bring our models to Qwen/Qwen3-4B-Thinking-2507 and the other 2507 Qwen 3 models - coming very soon!

We want to bring SV3, Esper 3, and our Experimental Reasoning finetunes to more models ASAP. Help us out: sequelbox/SupportOpenSource

Open source matters. Fight for it with us.

love,
allegra
sometimesanotion

This is a very cool release! I really enjoy the ShiningValiant series!

Do you see potential to prune experts or layers from the gpt-oss-20b model to downsize it, and then finetune?

·
sequelbox

thank you so much <3

yeah the particular combo that is oss-20b (larger experts + smaller amount of experts + already trained at MXFP4 so no easy gains from just-make-it-smaller-with-quantization-instead) seems well suited for this vs the qwen 3 30b-a3b type of MoE. definitely encourage this type of experimentation in general :)

In this post