nvidia · audio asr

parakeet-tdt-0.6b-v2

nvidia/parakeet-tdt-0.6b-v2

parakeet-tdt-0.6b-v2 at Q4_K_M is exactly 475,491,840 bytes (0.44 GiB / 0.48 GB) — an effective 6.157 bits per weight, not the nominal 4.

From the file· summed from 1 file(s)
Parameters
618M
Architecture
parakeet
Context
native (config.json)
License
cc-by-4.0

Shipped quantizations

exact bytes, summed from published files
QuantSizeExact bytesEffective bpwTensorsPublisher
Q4_K0.37 GiB396,937,9845.139cstr
Q4_K_M0.44 GiB475,491,8406.157697handy-computer
Q5_K_M0.50 GiB539,012,6086.979697handy-computer
Q6_K0.56 GiB600,408,5767.774697handy-computer
Q8_00.62 GiB664,466,3688.603725cstr
Q8_00.68 GiB729,574,9129.446697handy-computer
F161.15 GiB1,237,334,59216.021697handy-computer
F322.30 GiB2,471,550,27232.001handy-computer

Measured

published by a third party, attributed below
MetricValueWhat it means
rtf6038.1
RTFx6038.1Higher is better — audio seconds processed per second of compute.
Word error rate2.73%Lower is better — the share of words transcribed incorrectly.
Word error rate10.40%Lower is better — the share of words transcribed incorrectly.
Word error rate1.27%Lower is better — the share of words transcribed incorrectly.
Word error rate8.15%Lower is better — the share of words transcribed incorrectly.
Word error rate10.78%Lower is better — the share of words transcribed incorrectly.
Word error rate9.10%Lower is better — the share of words transcribed incorrectly.
Word error rate5.68%Lower is better — the share of words transcribed incorrectly.
Word error rate1.94%Lower is better — the share of words transcribed incorrectly.
Word error rate3.78%Lower is better — the share of words transcribed incorrectly.
Word error rate8.23%Lower is better — the share of words transcribed incorrectly.
Word error rate5.39%Lower is better — the share of words transcribed incorrectly.
Benchmarked· by open-asr-leaderboard-english-short-latest

RTFx measured by the Open ASR Leaderboard on a single datacenter GPU at a large batch size. It ranks models against each other; it says nothing about throughput on consumer hardware. We reproduce these figures with attribution; they are not ours and we have not verified the runs. Source: open-asr-leaderboard-english-short-latest.

Compare with

same modality, comparable size

Will it run on your card?

full quant x context sweep

Why other calculators give a different number

A parameters × bits ÷ 8 estimate puts Q4_K_M at roughly 0.32 GiB. The real file is 0.44 GiB, because a quantization is a mixture and some tensors are always kept at higher precision.

Architecture

Architecture unavailable — this repository is gated and no ungated mirror was found. Exact file sizes above are still authoritative; only the KV math needs the config.

Questions people ask

How much VRAM does parakeet-tdt-0.6b-v2 need?
Q4_K_M is exactly 475,491,840 bytes (0.44 GiB) in weights. Add the KV cache, which depends on your context length, plus roughly half a gigabyte of runtime overhead.
Which quantization of parakeet-tdt-0.6b-v2 should I use?
Q4_K_M is the usual default. Pick the largest quantization that fits your card at the context you actually need — the table above gives exact sizes for every one published.