model family
Whisper fine-tuned per Indian language (vasista22) logo

Whisper fine-tuned per Indian language (vasista22)

Published by vasista22. 4 members measured on real machines, across 4 builds. Every number below was signed by the machine that produced it — none of it is from a model card.

what it takes, what it returns
audio → text

Declared by the network’s catalog, not measured — a signature says what this family would be asked. Whether it answered is the signed capability flags on each member.

4 of 13 builds measured · 9 still open
what each build proved
hi-small
kn-small
ta-small
te-small
hi-small · Speech in — provedkn-small · Speech in — provedta-small · Speech in — provedte-small · Speech in — proved
Speech in
Whisper small hindi (vasista22)1 build measured →
vasista22/whisper-hindi-smallhi-small · vasista22 · speech-hf · 1 machine
—recalled
—served
—best tok/s
Whisper small kannada (vasista22)1 build measured →
vasista22/whisper-kannada-smallkn-small · vasista22 · speech-hf · 1 machine
—recalled
—served
—best tok/s
Whisper small tamil (vasista22)1 build measured →
vasista22/whisper-tamil-smallta-small · vasista22 · speech-hf · 1 machine
—recalled
—served
—best tok/s
Whisper small telugu (vasista22)1 build measured →
vasista22/whisper-telugu-smallte-small · vasista22 · speech-hf · 1 machine
—recalled
—served
—best tok/s

Builds this network knows of

Declared, not measured — this is what the catalog says exists and what it would cost a machine to hold. Nothing here is signed, and none of it claims the build works.

how much memory does your machine have

13 of 13 builds fit a 16 GB machine — budgeting 70% of it, or 11.2 GiB, because the cache and the runtime want the rest.

  • vasista22/whisper-gujarati-smallFP32speech-hfderived by vasista22
    0.9 GiB
  • vasista22/whisper-hindi-smallFP32speech-hfmeasured herederived by vasista22
    0.9 GiB
  • vasista22/whisper-kannada-smallFP32speech-hfmeasured herederived by vasista22
    0.9 GiB
  • vasista22/whisper-tamil-smallFP32speech-hfmeasured herederived by vasista22
    0.9 GiB
  • vasista22/whisper-telugu-smallFP32speech-hfmeasured herederived by vasista22
    0.9 GiB
  • vasista22/whisper-gujarati-mediumFP32speech-hfderived by vasista22
    2.8 GiB
  • vasista22/whisper-hindi-mediumFP32speech-hfderived by vasista22
    2.8 GiB
  • vasista22/whisper-kannada-mediumFP32speech-hfderived by vasista22
    2.8 GiB
  • vasista22/whisper-tamil-mediumFP32speech-hfderived by vasista22
    2.8 GiB
  • vasista22/whisper-telugu-mediumFP32speech-hfderived by vasista22
    2.8 GiB
  • vasista22/whisper-hindi-large-v2FP32speech-hfderived by vasista22
    5.8 GiB
  • vasista22/whisper-tamil-large-v2FP32speech-hfderived by vasista22
    5.8 GiB
  • vasista22/whisper-telugu-large-v2FP32speech-hfderived by vasista22
    5.8 GiB

Sizes are what the catalog declares, not what a machine measured — fitting on disk is necessary and not sufficient. Nobody here has run these on your machine, and this narrows the candidates rather than certifying one.

the same builds, by member
Whisper medium gujarati (vasista22)769M
vasista22/whisper-gujarati-mediumFP32speech-hf2.8 GiB on diskderived by vasista22
Whisper small gujarati (vasista22)244M
vasista22/whisper-gujarati-smallFP32speech-hf922 MiB on diskderived by vasista22
Whisper large-v2 hindi (vasista22)1.5B
vasista22/whisper-hindi-large-v2FP32speech-hf5.8 GiB on diskderived by vasista22
Whisper medium hindi (vasista22)769M
vasista22/whisper-hindi-mediumFP32speech-hf2.8 GiB on diskderived by vasista22
Whisper small hindi (vasista22)244M
vasista22/whisper-hindi-smallFP32speech-hf922 MiB on diskderived by vasista22
Whisper medium kannada (vasista22)769M
vasista22/whisper-kannada-mediumFP32speech-hf2.8 GiB on diskderived by vasista22
Whisper small kannada (vasista22)244M
vasista22/whisper-kannada-smallFP32speech-hf922 MiB on diskderived by vasista22
Whisper large-v2 tamil (vasista22)1.5B
vasista22/whisper-tamil-large-v2FP32speech-hf5.8 GiB on diskderived by vasista22
Whisper medium tamil (vasista22)769M
vasista22/whisper-tamil-mediumFP32speech-hf2.8 GiB on diskderived by vasista22
Whisper small tamil (vasista22)244M
vasista22/whisper-tamil-smallFP32speech-hf922 MiB on diskderived by vasista22
Whisper large-v2 telugu (vasista22)1.5B
vasista22/whisper-telugu-large-v2FP32speech-hf5.8 GiB on diskderived by vasista22
Whisper medium telugu (vasista22)769M
vasista22/whisper-telugu-mediumFP32speech-hf2.8 GiB on diskderived by vasista22
Whisper small telugu (vasista22)244M
vasista22/whisper-telugu-smallFP32speech-hf922 MiB on diskderived by vasista22
ask your own gram

Your concierge can answer this from the network rather than from this page. The ask carries the operator id, so it resolves to a capability and an engine — not to a name somebody has to recognise.

What has this network measured about the Whisper fine-tuned per Indian language (vasista22) family (epn.family.whisper-indic-ft), and can my gram run any of it? See /inferences/whisper-indic-ft

open a gramx with this ask →epn.family.whisper-indic-ft