Finetune of PleIAs/Baguettotron on the Dutch subset of the SYNTH dataset (1B tokens).
Also fixed the EOS token issue by finetuning on the correct token id (not super reliable yet).
Finetune of PleIAs/Baguettotron on the Dutch subset of the SYNTH dataset (1B tokens).
Also fixed the EOS token issue by finetuning on the correct token id (not super reliable yet).
model.safetensors 642 MBtraining_args.bin 5.9 kB有文件？提交种子。
提交种子