Audio Classification
PEFT
Safetensors
English
audio
audio-decision
qwen2_5_omni
lora
streaming
custom-code
research
Instructions to use TeenageWorks/Toddler-Jev with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use TeenageWorks/Toddler-Jev with PEFT:
Task type is invalid.
- Notebooks
- Google Colab
- Kaggle
Commit History
13 days ago
-
Add measured p50 and p90 inference latency
Verified signatureDocument streaming-window and 30-second-audio latencies on A100 40 GB BF16, including sample counts and timing scope.c052caa -
Center the banner and round its corners
Verified signature4610ddd -
Use Toddler-Jev banner as the model card title
Verified signatureAdd the supplied original banner and replace the text heading with it.30e5b65 -
Add model citation and update research attribution
Verified signatureList Vyom Nishant Patel as the sole researcher and citation author, identify Teenage Works Private Limited as publisher and affiliation, add its sourced registered office, and provide a bottom How to cite section with BibTeX.9a50548 -
Improve Toddler-Jev model card and usage guide
Verified signatureExplain audio candidate selection, application examples, inputs and outputs, local batch and sliding-window inference, training, architecture, and repository contents. Move required Qwen attribution into the licensing section. Documentation-only update.6c2941c -
Fix Hub adapter configuration parsing
Verified signatureOmit optional null task_type for the custom decision adapter, preserving PEFT default behavior. Apply the same serialization fix to fine-tuned exports. Weight files are unchanged.f6db0d7 -
Release Toddler-Jev 1.0.0
Verified signatureFinal trained LoRA and decision-head weights; pinned Qwen base and processor; local batched and windowed inference; LoRA and full fine-tuning. Built with Qwen.8e516d4 -
initial commit
Verified signature576d51e