Edge0-145B
AI overview
Sign in and the AI will write an overview from our coverage.
Headlines · 2
- Edge0-145B NAS Storage Test Shows 51x Slower Cold Reads
A workstation test read the same 105,368,210,944-byte file cold in 18.3 seconds from local NVMe and 15.6 minutes from a gigabit NAS, about 51 times slower. Based on the official model card cited by the author, the report recommends keeping the 51.20 GB lexicon layer on local SSD and the 40.08 GB expert library in system memory or fast local NVMe. The author did not download, load, or run Edge0-145B, so the report contains no measured inference throughput, latency, or memory use.
掘金 · 人工智能 · 🔥 0 - Edge0 Opens 145B and 120B On-Device Models
Edge0 open-sourced the Edge0-145B and Edge0-120B models, along with a unified inference framework and deployment tools. Edge0 describes Edge0-145B as a sparse model with 145 billion total parameters and about 10 billion active parameters. The company reports that it runs locally on compatible laptops with peak memory of 29GB and decoding above 80 tokens per second. A commentary report described hands-on tests of local 3D modeling, game and presentation workflows and noted that benchmark comparisons relied on Edge0-published results.
Hacker News · AI(100+ 分) · 🔥 3
Experience and discussion from the community
Share my Edge0-145B experienceAsk about Edge0-145B
Nobody has shared their experience with Edge0-145B yet.