III-tinystories / config.json
Naphon's picture
Upload III TinyStories model
98a5a1b verified
Raw History Blame Contribute Delete
932 Bytes
{
"architectures": [
"IIIForCausalLM"
],
"attention_bias": false,
"auto_map": {
"AutoConfig": "configuration_iii.IIIConfig",
"AutoModelForCausalLM": "modeling_iii.IIIForCausalLM"
},
"bos_token_id": 0,
"dtype": "float32",
"eos_token_id": 1,
"hidden_dim": 128,
"hidden_size": 128,
"intermediate_dim": 256,
"layer_dilation_size": [
1,
1,
2,
0
],
"layer_window_sizes": [
3,
5,
7,
0
],
"lora_dim": 64,
"max_position_embedding": 512,
"model_type": "iii",
"n_heads": 16,
"n_layers": 4,
"num_attention_heads": 16,
"num_hidden_layers": 4,
"pad_token_id": null,
"rms_norm_eps": 1e-06,
"rope_parameters": {
"rope_theta": 10000.0,
"rope_type": "default"
},
"rope_theta": 10000.0,
"tie_word_embedding": false,
"transformers_version": "5.14.1",
"use_cache": false,
"vocab_size": 1024
}