Text Generation
Transformers
Safetensors
jetmoe
alignment-handbook
generated_from_trainer
conversational
Inference Endpoints
jetmoe-8b-sft / config.json
YikangS's picture
update model config and readme
445846c
{
"_name_or_path": "jetmoe/jetmoe-8b-chat",
"model_type": "jetmoe",
"activation_function": "silu",
"architectures": [
"JetMoEForCausalLM"
],
"bos_token_id": 1,
"eos_token_id": 2,
"ffn_hidden_size": 5632,
"kv_channels": 128,
"layer_norm_epsilon": 1e-05,
"length_penalty": 1.0,
"moe_num_experts": 8,
"moe_top_k": 2,
"hidden_size": 2048,
"num_hidden_layers": 24,
"n_positions": 4096,
"num_attention_heads": 32,
"num_key_value_heads": 16,
"num_layers": 24,
"rms_norm_eps": 1e-05,
"rope_theta": 10000.0,
"rotary_percent": 1.0,
"tie_word_embeddings": true,
"transformers_version": null,
"use_cache": true,
"vocab_size": 32000,
"glu": true
}