hf-transformers-bot's picture
Update tiny models for PhimoeForCausalLM
e6e4cdd verified
Raw
History Blame Contribute Delete
924 Bytes
{
"architectures": [
"PhimoeForCausalLM"
],
"attention_bias": false,
"attention_dropout": 0.0,
"bos_token_id": 1,
"dtype": "float32",
"eos_token_id": 32000,
"hidden_act": "gelu",
"hidden_size": 32,
"initializer_range": 0.02,
"input_jitter_noise": 0.0,
"intermediate_size": 32,
"is_decoder": true,
"lm_head_bias": false,
"max_position_embeddings": 512,
"model_type": "phimoe",
"num_attention_heads": 2,
"num_experts_per_tok": 2,
"num_hidden_layers": 2,
"num_key_value_heads": 2,
"num_local_experts": 16,
"output_router_logits": false,
"pad_token_id": 32000,
"rms_norm_eps": 1e-05,
"rope_parameters": {
"rope_theta": 1000000.0,
"rope_type": "default"
},
"router_aux_loss_coef": 0.001,
"router_jitter_noise": 0.01,
"sliding_window": null,
"tie_word_embeddings": false,
"transformers_version": "5.16.0.dev0",
"use_cache": true,
"vocab_size": 32011
}