jdopensource/JoyAI-LLM-Flash-Base
2487
1{2 "architectures": [3 "DeepseekV3ForCausalLM"4 ],5 "attention_bias": false,6 "attention_dropout": 0.0,7 "auto_map": {8 "AutoConfig": "configuration_deepseek.DeepseekV3Config",9 "AutoModel": "modeling_deepseek.DeepseekV3Model",10 "AutoModelForCausalLM": "modeling_deepseek.DeepseekV3ForCausalLM"11 },12 "bos_token_id": 0,13 "eos_token_id": 1,14 "ep_size": 1,15 "first_k_dense_replace": 1,16 "hidden_act": "silu",17 "hidden_size": 2048,18 "initializer_range": 0.02,19 "intermediate_size": 7168,20 "kv_lora_rank": 512,21 "max_position_embeddings": 131072,22 "model_type": "joyai_llm_flash",23 "moe_intermediate_size": 768,24 "moe_layer_freq": 1,25 "n_group": 1,26 "n_routed_experts": 256,27 "n_shared_experts": 1,28 "norm_topk_prob": true,29 "num_attention_heads": 32,30 "num_experts_per_tok": 8,31 "num_hidden_layers": 40,32 "num_key_value_heads": 32,33 "num_nextn_predict_layers": 1,34 "q_lora_rank": 1536,35 "qk_nope_head_dim": 128,36 "qk_rope_head_dim": 64,37 "rms_norm_eps": 1e-06,38 "rope_theta": 32000000,39 "routed_scaling_factor": 2.5,40 "scoring_func": "sigmoid",41 "tie_word_embeddings": false,42 "topk_group": 1,43 "topk_method": "noaux_tc",44 "torch_dtype": "bfloat16",45 "transformers_version": "4.44.2",46 "use_cache": true,47 "v_head_dim": 128,48 "vocab_size": 12928049}50 