Modelpublic

JoaoZaokk/parakeet-tdt-1.1b-ggml

sourceHugging Facecc-by-4.0updated 15d agoView on Hugging Face

0likes

Model Card

parakeet-tdt-1.1b-ggml

GGML conversion of nvidia/parakeet-tdt-1.1b (.nemo) for the Parakeet TDT engine that ships inside whisper.cpp ≥ 1.9, in f16 plus q80 / q50 / q4_0.

Source checkpoint: nvidia/parakeet-tdt-1.1b by NVIDIA · License: cc-by-4.0 (unchanged; this repo only re-packages the weights) Engine: Load with whisper.cpp ≥ 1.9 parakeet-cli -m <file> (the Parakeet TDT engine that ships inside whisper.cpp). Not compatible with mudler/parakeet.cpp GGUF files.

Files

File	Quantization	Size
`ggml-parakeet-tdt-1.1b-f16.bin`	f16	2145 MB
`ggml-parakeet-tdt-1.1b-q4_0.bin`	q4_0	609 MB
`ggml-parakeet-tdt-1.1b-q5_0.bin`	q5_0	743 MB
`ggml-parakeet-tdt-1.1b-q8_0.bin`	q8_0	1143 MB

f16 is the lossless conversion; q8_0 is nearly identical in accuracy at ~55 % of the size; q5_0/q5_k are the phone-friendly choice; q4_* is smallest with a small accuracy cost.

How these were made

Converted from the upstream checkpoint with the engine's own converter, then quantized with the engine's quantizer. Each variant was checked by transcribing short Portuguese and English samples before upload.

Attribution

Weights are derivative works of the upstream model and keep its license. Please cite the original authors (NVIDIA). The conversion and hosting here are maintained by JoaoZaokk so that the download links used by the Odysseus / Open WebUI native apps stay stable. No warranty.