Barist.7 / tokenizer_config.json
KucLab's picture
Standard LlamaForSequenceClassification layout: loads with plain transformers, no trust_remote_code
9e9e04f verified
Raw History Blame Contribute Delete
156 Bytes
{
"tokenizer_class": "PreTrainedTokenizerFast",
"pad_token": "[PAD]",
"bos_token": "[BOS]",
"eos_token": "[ASSISTANT]",
"model_max_length": 4096
}