Skip to content

laya:multilingual

10 TagsUpdated 322M params1024 context100+ languagesApache-2.0by Convai Innovations

100+ languages, 1024-token context; up to ~2.2× faster on batched calls.

multilingual322m
ollaya run laya:multilingual --preset triage "I was charged twice for my subscription this month and want a refund."

Details

  • graphe44370875133 · 3 MBonnx · mmBERT-base · 322M · fp32
  • graph236fbdec33bd · 3 MBonnx · mmBERT-base · 322M · fp16
  • weights9d628fd971b7 · 644 MBhuggingface.co/convaiinnovations/laya/resolve/aa8c91c…/multilingual/model.safetensors
  • tokenizer609d8f4c067c · 34 MBhuggingface.co/convaiinnovations/laya/resolve/aa8c91c…/multilingual/tokenizer/tokenizer.json
  • decision5376e3133945 · 455 B{"engine": "onnx", "family": "laya", "encoder": "jhu-clsp/mmBERT-base", "layout": "laya-markers-v1", …}
  • calibration316915b3e458 · 84 B{"temperature": [1.0, 1.0, 1.0]}
  • licenseec56c3e69fab · 10 KBLaya by Convai Innovations (https://huggingface.co/convaiinnovations/laya)

Every layer is checked against its sha256 when it is pulled. Weights and tokenizers download from the model author's Hugging Face repository at a pinned commit; Ollaya never re-hosts them.

Precision variants

Laya readme and all models