Prometh-MOEM-32-tiny / config.json
Wtzwho's picture
Initial commit with model files
b518f38
{
"_name_or_path": "TinyLlama/TinyLlama-1.1B-Chat-v1.0",
"architectures": [
"MixtralForCausalLM"
],
"attention_bias": false,
"attention_dropout": 0.0,
"bos_token_id": 1,
"eos_token_id": 2,
"hidden_act": "silu",
"hidden_size": 2048,
"initializer_range": 0.02,
"intermediate_size": 5632,
"max_position_embeddings": 2048,
"model_type": "mixtral",
"num_attention_heads": 32,
"num_experts_per_tok": 2,
"num_hidden_layers": 22,
"num_key_value_heads": 4,
"num_local_experts": 32,
"output_router_logits": false,
"pretraining_tp": 1,
"rms_norm_eps": 1e-05,
"rope_scaling": null,
"rope_theta": 10000.0,
"router_aux_loss_coef": 0.001,
"sliding_window": null,
"tie_word_embeddings": false,
"torch_dtype": "bfloat16",
"transformers_version": "4.37.2",
"use_cache": true,
"vocab_size": 32000
}