bielik-minitron-jev-v0.3

bielik-minitron-jev-v0.3

A Jev-style System One decision model made with any2jev from speakleash/Bielik-Minitron-7B-v3.0-Instruct. State in, typed Choice / Score / Noul answers with calibrated probabilities out, in one forward pass. No text is generated.

pip install "any2jev[serve]"
any2jev serve hf://wwydmanski/bielik-minitron-jev-v0.3      # POST /v1/systemone, TypeSafe SDK compatible
any2jev ask hf://wwydmanski/bielik-minitron-jev-v0.3 --state "My payouts have failed 3 days in a row, fix this ASAP" \
    --choice "Which team? | billing, technical, sales" --noul "Is this urgent?"

What is in this repo

• adapter/: LoRA adapter (r=64) on speakleash/Bielik-Minitron-7B-v3.0-Instruct, trained with the vocabulary head removed
• head.safetensors: the pointer head (dim 256) that scores options against the decision token
• any2jev.json: delimiters, mode (packed), temperature 1.282 fitted on validation
• tokenizer/: the base tokenizer (delimiter tokens reused or added)
• train_report.json, eval.json: training config, history and held-out metrics

Training

• data: data/train_mix3v01.jsonl; 0.0 M trainable parameters, 1.0 epoch(s), lr 3e-05, batch 1 x 4
• wall clock: 181 min on one consumer GPU

Limitations

Trained on a few thousand labelled decisions from a handful of sources; expect the accuracy above on similar inputs and lower accuracy off-distribution. Probabilities are calibrated on the validation split, not a guarantee per answer. Keep arithmetic, dates and counting in code, as Jev's own docs recommend.

Independent project, not affiliated with TypeSafe AI. Apache-2.0.