Changes / Feeds / openai/whisper-large-v3
openai/whisper-large-v3
2 changes to a tracked key since 2023-11-07, newest first. Head as last read: 06f233fe06e710322aca913c1bc4249a0d71fce1.
GPU memory to fine-tune it, card by card: whisper-large-v3 on /m/.
| Date | Advisory | What changed, and what it does |
|---|---|---|
| 2023-11-16 | x_NXMC-2023-0031 breaking | EOS changed from 50256 to 50257; anything fine-tuned or served against the old value stops at a different token; BOS changed from 50256 to 50257; prompts built by the old tokenizer start with a different token (2023-11-16).ca3e273640c8 -> 9ef08b92d511 |
| 2023-11-07 | x_NXMC-2023-0030 | pad token unset to "<|endoftext|>"; batched training masks a different id (2023-11-07).8a2a8a341931 -> d8411bd4e55c |
Pin it, and let a bot move the pin
from_pretrained("openai/whisper-large-v3", revision="06f233fe06e710322aca913c1bc4249a0d71fce1")With the Renovate preset, a pull request moves this pin when the head moves, and links back here.
[](https://www.neurarch.com/changes-feeds/models/openai/whisper-large-v3.html)