An open-weights System One model from Convai Innovations. A fully fine-tuned ModernBERT-large encoder with a from-scratch decision head that scores one marker per option and answers every question in a single 33–39 ms pass. Runs on your own CPU or GPU.
19 files · 2.2 GBMirrored from https://huggingface.co/convaiinnovations/laya
email_utils.py1b1c0a63.8 kB
model.safetensors891102d804 MB
rl_agent_api.pybe3b4684.6 kB
rl_agent_config.jsonae287b5745 B
rl_common.py8d8361119 kB
config.jsonbf3ab802.0 kB
results.json13b0c982.9 kB
encoder/config.json83f69161.9 kB
model.safetensors9d628fd614 MB
rl_agent_config.json2506173472 B
tokenizer/tokenizer_config.json6c6b2d8524 B
tokenizer/tokenizer.json609d8f433 MB
tokenizer_config.json50044de308 B
tokenizer.json6c8aaa93.4 MB
encoder/config.json5268d242.0 kB
model.safetensors4fa56de804 MB
rl_agent_config.jsonebf0cd5847 B
tokenizer/tokenizer_config.json08d4cf3337 B
tokenizer/tokenizer.json6c8aaa93.4 MB
Every file resolves through a stable path — /v1/models/convai-innovations/laya/resolve/2026.09.25/<path> — while the object underneath is content-addressed and shared with every other version that holds the same bytes. API reference.