4 1 week ago

Arc is a small, experimental English model from Geocentric (120.8M parameters, 1,024-token context, 483 MB) for co-writing: brainstorming phrasing and drafting short pieces. Runs offline on almost any machine. It can invent facts. Apache 2.0.

e3d3acbf1b48 ยท 435B
{
"tokenizer_class": "PreTrainedTokenizerFast",
"model_max_length": 1024,
"chat_template": "{% for m in messages %}{% set role = 'user' if m['role'] == 'tool' else m['role'] %}<|{{ role }}|>\n{{ m['content'] }}<|eot|>\n{% endfor %}{% if add_generation_prompt %}<|assistant|>\n{% endif %}",
"clean_up_tokenization_spaces": false,
"bos_token": "<bos>",
"eos_token": "<|eot|>",
"pad_token": "<pad>",
"unk_token": "<unk>"
}