Hugging Face's logo
Carlo Moro's picture
šŸ‘‹ Open to Work

Carlo Moro PRO

cnmoro
55 29 616
Gargaz's profile picture divyabch-auhan1993's profile picture Bruno's profile picture
Ā·

AI & ML interests

Retrieval & Small Models & PTBR Resources

Recent Activity

liked a model 2 days ago
LiquidAI/d1-3B-GGUF
liked a model 3 days ago
Cactus-Compute/whistle
reacted to SeaWolf-AI's post with ā¤ļø 7 days ago
šŸ’» Data-center AI, now on a laptop: POCKET-Darwin-180B We're releasing a 4-bit GGUF build of Darwin-180B-RSI, #1 on seven official Hugging Face leaderboards (self-reported), that runs without a GPU. šŸ“¦ 360 GB ā†’ 111 GB (4-bit GGUF, 4 files) šŸ–„ļø No GPU: one server CPU (16 threads) at 18.4ā€“21.0 tokens/s šŸ’» RTX 5060 laptop (8 GB VRAM) + 32 GB RAM: 4.17 tokens/s šŸ§Š 128 GB mini PC: whole model in memory, no GPU needed šŸŽÆ MMLU-Pro, 2,000 questions, paired: original 87.65% = 4-bit 87.65% How? Ā· Only ~3B of 180B parameters are active per token (10 of 512 experts) Ā· llama.cpp streams just the needed experts from SSD, so 32 GB RAM is enough Ā· Graft quantization: we took the proven Unsloth UD-Q4_K_XL base build and swapped in only the 300 tensors our RSI training changed (300/300 verified) Under the hood is Model-level Recursive Self-Improvement. The model solves verifiable problems, keeps only its own solutions that check out as correct, and trains on them. No human-written solutions or reasoning traces. Built for teams that can't send data to an external cloud (defense, finance, public sector) to run a top-tier model fully offline. šŸ“ Article: https://huggingface.co/blog/FINAL-Bench/data-center-ai-now-on-a-laptop-pocket-darwin-180b šŸ¤— Model: https://huggingface.co/FINAL-Bench/POCKET-Darwin-180B-GGUF šŸ§¬ Original: https://huggingface.co/FINAL-Bench/Darwin-180B-RSI #Darwin #RSI #GGUF #llamacpp #OnDevice #MoE
View all activity

Organizations

Wise Intelligence's profile picture