Hugging Face's logo

AdaptiveChunking
/
pilot_Bpc_s0

PyTorch
Safetensors
dynamic-chunking
tokenization
hnet
byte-level
beetle-hnet
legacy-trainer
Model card Files
xet
 Community
pilot_Bpc_s0 / README.md
suchirsalhan's picture
suchirsalhan
main <- step 6000
7d40208 verified
preview code
|
Raw History Blame Contribute Delete
1.03 kB
metadata 
license: apache-2.0
library_name: pytorch
tags:
  - dynamic-chunking
  - tokenization
  - hnet
  - byte-level
  - beetle-hnet
  - legacy-trainer

pilot_Bpc_s0

Byte-level H-Net run from beetle-hnet, trained with the legacy trainer (legacy/tier1/train.py). Split out of the bundled repo AdaptiveChunking/hnet-chunking-pilots into the Beetle per-run layout.

  • step-N branches: 1 retained checkpoints (6000), each with model.safetensors and config.json (model_config, training config and the tied_weights map: restore with sd[dst] = sd[src] for every entry).
  • main: the latest weights plus final.json and log.jsonl.
  • The legacy trainer kept no optimiser state, so there is no training_state/ here. FLORES+ boundary probes for this run are in the dataset AdaptiveChunking/hnet-chunking-probes under pilot_Bpc_s0/.