Parakeet GGUF — models for parakeet.cpp GGUF format weights for parakeet.cpp, a C++/ggml port of NVIDIA NeMo Parakeet that matches the upstream PyTorch models on CPU. This single repo collects every supported model × quantization as a flat set of .gguf files — download just the one you need. F16 is the recommended default — same accuracy as F32, ~1.7× smaller, and typically the fastest on modern CPUs via ggml's F32×F16 matmul fast path. Models tdt ctc 110m Source: nvidia/parakeet tdt ctc 110m · Hybrid TDT+CTC (FastConformer) · heads: TDT + CTC File Variant Size WER vs NeMo : : tdt ctc 110m f16.gguf ← recommended F16 267.5 MB 0.0000 tdt ctc 110m q8 0.gguf Q8 0 177.8 MB 0.0000 tdt ctc 110m q6 k.gguf Q6 K 155.9 MB not measured tdt ctc 110m q5 k.gguf Q5 K 143.3 MB not measured tdt ctc 110m q4 k.gguf Q4 K 131.4 MB 0.0000 realtime eou 120m v1 Source: nvidia/parakeet realtime eou 120m v1 · Cache aware streaming RNNT (FastConformer, EOU/EOB) · heads: RNNT (streaming) File Variant Size WER vs NeMo : : realtime eou 120m v1 f16.gguf ← recommended F16 266.5 MB not measured realtime eou 120m v1 q8 0.gguf Q8 0 176.0 MB not measured realtime eou 120m v1 q6 k.gguf Q6 K 153.9 MB not measured realti…
We use cookies for essential functionality and analytics. You can accept or reject analytics cookies.Cookie policy