LifeMate models

Checkpoints for the code in https://github.com/lducc/lifemate (private).

Path	What
`llm_adapter/`	LoRA r32 for Qwen/Qwen3.5-4B (`sft_a0`, 3 epochs, final step)
`asr/avg2-4.pt`	Gipformer 1.5 (g-group-ai-lab/gipformer1.5-68M-rnnt) fine-tuned on all train audio, epochs 2-4 averaged
`asr/public.jsonl`	that model's transcripts of public_test audio; put it at `runs/asr/gipall_public.jsonl`

hf download lducc/lifemate-models --local-dir models/lifemate
python -m src.predict a0 gipall_public <adapter> public

src.predict reads the adapter from lducc/lifemate-adapters; point ADAPTERS in src/predict.py at this repo and pass llm_adapter as the adapter.

Downloads last month: -; Downloads are not tracked for this model. How to track

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support