MusicGen Medium โ€” INT8 (ONNX)

Quantized ONNX export of facebook/musicgen-medium for on-device inference.

Quantization Details

Property Value
Base Model facebook/musicgen-medium
Format ONNX
Quantization Dynamic INT8 quantization (per-channel)
Export Tool Optimum + ONNXRuntime

Model Components

File Description
text_encoder.onnx T5 text encoder
decoder_model.onnx MusicGen decoder (first pass)
decoder_with_past_model.onnx MusicGen decoder (with KV cache)
encodec_decode.onnx EnCodec audio decoder
build_delay_pattern_mask.onnx Delay pattern mask builder

Usage

Designed for on-device inference with ONNX Runtime (Android/iOS/Desktop).

License

This model inherits the CC-BY-NC-4.0 license from the original MusicGen model.

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. ๐Ÿ™‹ Ask for provider support

Model tree for Ocean82/musicgen-medium-int8-onnx

Quantized
(2)
this model