john smith's picture

john smith

k249

AI & ML interests

None yet

Recent Activity

liked a model about 7 hours ago
intellectlabs/Kepler-8B-Instruct-v2
reacted to SeaWolf-AI's post with โค๏ธ about 7 hours ago
A small gift for anyone building or studying foundation models. Most "open" models hand you the weights and stop there. With Aether-7B-5Attn we wanted to hand over the whole thing โ€” so you can actually learn from it, reproduce it, and build on it: the data recipe, the training code, every hyperparameter, the complete logs, and the intermediate checkpoints. All Apache-2.0, reproducible byte-for-byte. What you can do with it: ๐Ÿ” Rebuild it from scratch, or fork the recipe for your own model ๐Ÿ”ฌ Study a real heterogeneous-attention MoE โ€” 49 layers place 5 attention mechanisms on a 7ร—7 Latin square, arranged as a clean, attributable ablation ๐Ÿ“ˆ Trace training dynamics across the released checkpoints (110k / 115k / 162k) It's a modest 6.59B model, and an honest one โ€” the limitations (no KV-cache in this build, small scale) are written right in the card. We're not claiming it's special. If any piece of it saves you time or teaches you something, that's exactly what we hoped for. ๐Ÿค— ๐Ÿ“– Full write-up โ†’ [blog] ยท https://huggingface.co/blog/FINAL-Bench/opensource-llm ๐Ÿ“ฆ Base ยท https://huggingface.co/FINAL-Bench/Aether-7B-5Attn ๐ŸŽฏ Instruct ยท https://huggingface.co/FINAL-Bench/Aether-7B-5Attn-it ๐Ÿš€ Live demo ยท https://huggingface.co/spaces/FINAL-Bench/Aether-Sovereign-AI ๐Ÿงฌ Collection ยท https://huggingface.co/collections/FINAL-Bench/aether-foundation-model #opensource #LLM #MoE #reproducibility #Apache2
liked a dataset about 7 hours ago
awacke1/DatasetOfDatasetsUSA
View all activity

Organizations

None yet