Buckets:
metadata
license: apache-2.0
library_name: pytorch
tags:
- rwkv
- rwkv-7
- language-model
- pretraining
Winner-RWKV-7-285M-20260601
RWKV-7 checkpoint trained in Jellyfish042/GatedDeltaNet-2.
- Architecture: RWKV-7 TimeMix,
rwkv7_time_1024x12 - Parameters printed by the training script:
219,611,136 - Pretraining data: globally shuffled FineWeb-Edu packed 100B-token run
- Sequence length:
4096 - Global batch size:
1024 - Seed:
3407 - Checkpoint file:
pytorch_model.bin
This is a project checkpoint from pretrain.py, not a Hugging Face Transformers checkpoint.
lm-eval-harness snapshot
| Task | Metric | Value |
|---|---|---|
lambada_openai |
acc | 0.3728 |
lambada_openai |
perplexity | 25.6649 |
wikitext |
word_perplexity | 24.0226 |
wikitext |
byte_perplexity | 1.8121 |
wikitext |
bits_per_byte | 0.8577 |
Xet Storage Details
- Size:
- 878 Bytes
- Xet hash:
- 40d9a407271be2d0790e4d844d299ba0458497a07e71bb144a7c875250068fb1
·
Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.