Beta
Explore
Marketplace
Neural Labs
Chat
Wallet
Docs
rinn-taiko-comba-180m-ja-base-v1 – AI Model by WariHima | AlphaNeural AI
You can deploy this model and start earning money today!
WariHima
/
rinn-taiko-comba-180m-ja-base-v1
like
0
safetensors
comba
custom_code
ja
cc-by-sa-4.0
us
Views
No views yet
Model card
Files and Versions
Community
API
Deploy
隣大河 comba 180m base ja v1
architecture:
comba(fla)
tarined by (日本語):
rtx3060 * 2.5h * 6
トレーニングコードは ./training_codesにあります。
学習の際はflash-linear-attentionライブラリのインストールとtrust_remote_codeの有効化が必要です。
tokenizer from ku-nlp/gpt2-small-japanese-char
+(qwne3 chat format | special tokens)
dataset:
OmniAICreator/Japanese-Wikipedia-202506
use 150k rows 1 epoch * 6, = 900k row (512 seq)