hf-transformers-bot's picture
Update tiny models for YoutuForCausalLM
772b312 verified
Raw
History Blame Contribute Delete
916 Bytes
{
"architectures": [
"YoutuForCausalLM"
],
"attention_bias": false,
"attention_dropout": 0.0,
"bos_token_id": 128000,
"dtype": "float32",
"embedding_initializer_range": 0.04,
"eos_token_id": 128001,
"head_dim": 32,
"hidden_act": "gelu",
"hidden_size": 32,
"initializer_range": 0.02,
"intermediate_size": 32,
"is_decoder": true,
"kv_lora_rank": 16,
"max_position_embeddings": 512,
"model_type": "youtu",
"num_attention_heads": 2,
"num_hidden_layers": 2,
"num_key_value_heads": 2,
"pad_token_id": 128001,
"q_lora_rank": 32,
"qk_head_dim": 64,
"qk_nope_head_dim": 32,
"qk_rope_head_dim": 32,
"rms_norm_eps": 1e-06,
"rope_interleave": true,
"rope_parameters": {
"rope_theta": 10000.0,
"rope_type": "default"
},
"tie_word_embeddings": false,
"transformers_version": "5.16.0.dev0",
"use_cache": true,
"v_head_dim": 32,
"vocab_size": 128256
}