calm3-22b-RP-v2 / config.json
Aratako's picture
Trained with Unsloth
3ae58c7 verified
raw
history blame
766 Bytes
{
"_name_or_path": "cyberagent/calm3-22b-chat",
"architectures": [
"LlamaForCausalLM"
],
"attention_bias": false,
"attention_dropout": 0.0,
"bos_token_id": 0,
"eos_token_id": 65001,
"hidden_act": "silu",
"hidden_size": 6144,
"initializer_range": 0.02,
"intermediate_size": 16384,
"max_position_embeddings": 16384,
"mlp_bias": false,
"model_type": "llama",
"num_attention_heads": 48,
"num_hidden_layers": 48,
"num_key_value_heads": 48,
"pad_token_id": 1,
"pretraining_tp": 1,
"rms_norm_eps": 1e-05,
"rope_scaling": null,
"rope_theta": 1000000.0,
"tie_word_embeddings": false,
"torch_dtype": "bfloat16",
"transformers_version": "4.44.2",
"unsloth_version": "2024.8",
"use_cache": true,
"vocab_size": 65024
}