Uchaka PRO

Babu420

AI & ML interests

None yet

Recent Activity

reacted to evalstate's post with 🔥 1 day ago

Hugging Face MCP Server v0.2.40 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Improved progressive disclosure and descriptions for Jobs tool.

reacted to AdinaY's post with 🚀 3 days ago

Kimi K2 Thinking is now live on the hub 🔥 https://huggingface.co/moonshotai/Kimi-K2-Thinking ✨ 1T MoE for deep reasoning & tool use ✨ Native INT4 quantization = 2× faster inference ✨ 256K context window ✨ Modified MIT license

updated a model 3 days ago

Babu420/aquif-moe-800M-Q4_K_M-GGUF

View all activity

Organizations

None yet

reacted to evalstate's post with 🔥 1 day ago

Post

1865

Hugging Face MCP Server v0.2.40
~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Improved progressive disclosure and descriptions for Jobs tool.

reacted to AdinaY's post with 🚀 3 days ago

Post

2533

Kimi K2 Thinking is now live on the hub 🔥

moonshotai/Kimi-K2-Thinking

✨ 1T MoE for deep reasoning & tool use
✨ Native INT4 quantization = 2× faster inference
✨ 256K context window
✨ Modified MIT license

updated a model 3 days ago

Babu420/aquif-moe-800M-Q4_K_M-GGUF

Text Generation • 3B • Updated 3 days ago • 17

published a model 3 days ago

Babu420/aquif-moe-800M-Q4_K_M-GGUF

Text Generation • 3B • Updated 3 days ago • 17

reacted to DheemanthReddy's post with ❤️ 5 days ago

Post

1086

We just released Maya-1-Voice, an open source voice AI model with voice design and emotions.

Describe voices in natural language. Add 20+ emotions like <laugh>, <cry>, <whisper> inline. 3B parameters, production-ready, runs on single GPU with vLLM.

Apache 2.0. Built on Llama backbone, predicts SNAC codec tokens for real-time streaming.

Model: https://huggingface.co/maya-research/maya-1-voice

reacted to codelion's post with 🔥 5 days ago

Post

3533

On this day in 2019, OpenAI released the final GPT-2 model as part of their staged release. I still remember that November well - so much was happening, but GPT-2's release felt like a watershed moment for the field. It showed us what was possible with carefully trained language models.

To recreate some of that GPT-2 magic, I recently tackled an interesting challenge: can you pretrain a language model with just 1 billion tokens - roughly 1/10th of what GPT-2 used - and still get comparable performance? After 50+ systematic experiments testing different dataset mixtures, the answer is yes.

The result is codelion/gpt-2-70m, which achieves over 90% of GPT-2's benchmark performance despite being trained on 10x less data. The key was finding the optimal dataset composition: 50% high-quality textbook PDFs, 30% filtered web content, and 20% educational resources. It even beats GPT-2 on TruthfulQA (47.31% vs 40.69%).

If you're interested in the full story of how we discovered this optimal mixture and why curriculum learning catastrophically failed, check out the complete article: https://huggingface.co/blog/codelion/optimal-dataset-mixing

Sometimes less really is more - when you mix it right.