t.d.a.g. PRO

sequelbox

105 32 362

sequelbox.bsky.social

AI & ML interests

open source, infinite games. (they/them)

Recent Activity

liked a model 2 days ago

ertghiu256/Qwen3-1.7B-tiny-orchestrator

liked a model 7 days ago

mradermacher/Qwen3.6-27B-Esper4-i1-GGUF

liked a model 7 days ago

mradermacher/Qwen3.6-27B-Esper4-GGUF

View all activity

Organizations

liked a model 2 days ago

ertghiu256/Qwen3-1.7B-tiny-orchestrator

Text Generation • 2B • Updated 5 days ago • 67 • 2

liked 2 models 7 days ago

mradermacher/Qwen3.6-27B-Esper4-i1-GGUF

27B • Updated 7 days ago • 854 • 1

mradermacher/Qwen3.6-27B-Esper4-GGUF

27B • Updated 7 days ago • 499 • 1

replied to their post 7 days ago

GGUFs are now available:

https://huggingface.co/mradermacher/Qwen3.6-27B-Esper4-i1-GGUF
https://huggingface.co/mradermacher/Qwen3.6-27B-Esper4-GGUF

more to come very soon!

New activity in mradermacher/model_requests 7 days ago

https://huggingface.co/ValiantLabs/Qwen3.6-27B-Esper4

#2604 opened 8 days ago by

sequelbox

replied to their post 7 days ago

thank you! the Esper 4 mix is interesting because the datasets obviously have a lot of fundamental overlap; there are plenty of code tasks in Titanium/Mitakihara that could reasonably be in the Tachibana dataset, and some in Tachibana that could have gone the other way. by including the ~25% mix of 'old' queries in Titanium and Mitakihara it provides some balance here (many of the old Mitakihara queries in particular are chat/knowledge queries 'about' AI and AI-adjacent topics instead of specific tasks for the AI to perform) while still being highly relevant to Esper users.

have fun building :)

updated a Space 8 days ago

SupportOpenSource

😻

Support our open source releases!

updated a collection 8 days ago

Reasoning Datasets

Collection

Synthetic datasets generated using reasoning models, primarily the Deepseek-R1 and Deepseek-V3 series. • 16 items • Updated 8 days ago • 6

posted an update 8 days ago

Post

7361

NEW RELEASE: Esper 4 is here for Qwen 3.6 27b, along with our new datasets!

- NEW DATASET: Titanium 4 maximizes DevOps and architecture helpfulness, powered by high-difficulty agentic-focused DevOps and architecture data generated with DeepSeek-V4-Pro!
- NEW DATASET: Mitakihara 2 brings AI coding and expertise data for AI development, research, deployment, interpretability, operation and experimentation!
- Improved coding performance: challenging agentic coding queries from Tachibana 4 allow Esper 4 to tackle harder coding tasks across a variety of languages!

GET ESPER 4: ValiantLabs/Qwen3.6-27B-Esper4

Get the datasets for your own training:
sequelbox/Titanium4-DeepSeek-V4-Pro
sequelbox/Mitakihara2-DeepSeek-V4-Pro
sequelbox/Tachibana4-DeepSeek-V4-Pro

We've been working hard on Esper 4 - it's so exciting to finally bring it to everyone! We hope it helps you build.

We'll be expanding Esper 4 to more models as funding allows - donate for more, faster, better models and datasets: sequelbox/SupportOpenSource

The revolution is coming - we're here to fight for AI you can use and build on your own computer, not a giant corporation charging you for access at their discretion. We've seen what OpenAI, Anthropic, and the ultra-rich taking charge of the AI future looks like, and it's already very clear you won't like living in it. Choose a different future while you still can.

Open source must win.

More to come soon!

love, always,
allegra