0.7B-Bilingual-ChatML / config.json
aixk's picture
Upload scratch-built 0.7B bilingual model and 4-bit GGUF quantization
9fa4803 verified
raw
history blame contribute delete
302 Bytes
{
"architectures": [
"LlamaForCausalLM"
],
"bos_token_id": 1,
"eos_token_id": 2,
"hidden_size": 1536,
"intermediate_size": 4096,
"num_attention_heads": 12,
"num_hidden_layers": 24,
"vocab_size": 32002,
"model_type": "llama",
"rms_norm_eps": 1e-06,
"torch_dtype": "float32"
}