Text Generation
Safetensors
English
gpt2
dataset-mixing
pretraining
Eval Results (legacy)
gpt-2-70m / special_tokens_map.json
codelion's picture
Upload GPT-2 70M model trained with 40-30-30 dataset mixing (1B tokens)
9fac7e5 verified
Raw
History Blame Contribute Delete
99 Bytes
{
"bos_token": "<|endoftext|>",
"eos_token": "<|endoftext|>",
"unk_token": "<|endoftext|>"
}