1.53 GB

Ctrl+K

1 contributor

2.1-generation update: two-stage retrain (verbatimize/continuation release recipe + timing supervision on per-model top-10 discovered alignment heads). Fixes early-EOT / single-speaker longform failures.

f011e4a verified 18 days ago

.gitattributes

1.52 kB
Squash history to remove training artifacts from history about 2 months ago
LICENSE.md

17.8 kB
license: update commercial-licensing contact email to licensing@nyra-labs.com about 1 month ago
README.md

11 kB
Model card: sync README from nyrahealth/CrisperWhisper about 1 month ago
added_tokens.json

35.3 kB
Squash history to remove training artifacts from history about 2 months ago
config.json

1.35 kB
Squash history to remove training artifacts from history about 2 months ago
generation_config.json

3.86 kB
2.1-generation update: two-stage retrain (verbatimize/continuation release recipe + timing supervision on per-model top-10 discovered alignment heads). Fixes early-EOT / single-speaker longform failures. 18 days ago
merges.txt

494 kB
Squash history to remove training artifacts from history about 2 months ago
model.safetensors

1.53 GB
xet

2.1-generation update: two-stage retrain (verbatimize/continuation release recipe + timing supervision on per-model top-10 discovered alignment heads). Fixes early-EOT / single-speaker longform failures. 18 days ago
normalizer.json

52.7 kB
Squash history to remove training artifacts from history about 2 months ago
preprocessor_config.json

356 Bytes
Squash history to remove training artifacts from history about 2 months ago
special_tokens_map.json

2.19 kB
Squash history to remove training artifacts from history about 2 months ago
tokenizer.json

3.94 MB
Squash history to remove training artifacts from history about 2 months ago
tokenizer_config.json

288 kB
Squash history to remove training artifacts from history about 2 months ago
vocab.json

836 kB
Squash history to remove training artifacts from history about 2 months ago