yb2698 and Nathan (nathanhabib1011) released LoRA adapters on LFM2.5-24B for the open Pi harness and Claude Code. Nathan says the Pi adapter was RL-trained with TRL's AsyncGRPO in about 24 hours on one GPU and solves as many tasks with 39% fewer tool...
VQV Signal
Liquid AI's LFM2.5-24B gets LoRA adapters that cut agent tool calls by about 40%
yb2698 and Nathan (nathanhabib1011) released LoRA adapters on LFM2.5-24B for the open Pi harness and Claude Code. Nathan says the Pi adapter was RL-trained with TRL's AsyncGRPO in about 24 hours on one GPU and solves as many tasks with 39% fewer tool...
yb2698 and Nathan (nathanhabib1011) released LoRA adapters on LFM2.5-24B for the open Pi harness and Claude Code. Nathan says the Pi adapter was RL-trained with TRL's AsyncGRPO in about 24 hours on one GPU and solves as many tasks with 39% fewer tool calls, and Quentin Gallouédec echoed the result. Original post...
Hardware and robotics watchers may want to track whether this becomes a product, benchmark, or deployment signal.
VQV organizes public signals from inspectable sources. It does not independently verify the underlying report.
Signal Strength reflects source quality, relevance, freshness and evidence. Public Interest helps organize discovery; it is not proof of truth.
Public Interest components
VQV surfaced this signal because it is recent, relevant to AI Coding Tools, connected to Scobleizer News Pulse (X).
No login, cookies, social SDKs, or automatic posting.