← Back to all builds
Related builds
laya-apple
Same Laya weights on MLX GPU and Apple Neural Engine at once. P99 1593ms to 80ms on M4 Max.
Halo
Post-training framework for open-source models. Up to 2.8x TRL throughput, less peak memory, weights stay HuggingFace-native.
openJev Verdict
Open 151M decision model on ModernBERT. JevBench #2 on public tasks after the inference-engine fix. Weights and a WebGPU playground.
Laya
Open Apache 2.0 System One model. ModernBERT, typed choice/score/noul in one pass. Fine-tune notebook on Kaggle T4s. Not Jev.
