tiny-random/hy-v3
1218
1{2 "architectures": [3 "HYV3ForCausalLM"4 ],5 "attention_bias": false,6 "attention_dropout": 0.0,7 "bos_token_id": 120000,8 "dtype": "bfloat16",9 "enable_attention_fp32_softmax": false,10 "enable_lm_head_fp32": true,11 "enable_moe_fp32_combine": false,12 "eod_token_id": 120026,13 "eos_token_id": 120025,14 "expert_hidden_dim": 32,15 "first_k_dense_replace": 1,16 "head_dim": 32,17 "hidden_act": "silu",18 "hidden_size": 8,19 "initializer_range": 0.006,20 "intermediate_size": 32,21 "max_position_embeddings": 262144,22 "mlp_bias": false,23 "mlp_layer_types": [24 "dense",25 "sparse",26 "sparse",27 "sparse"28 ],29 "model_type": "hy_v3",30 "moe_intermediate_size": 32,31 "moe_router_enable_expert_bias": true,32 "moe_router_use_sigmoid": true,33 "num_attention_heads": 8,34 "num_experts": 192,35 "num_experts_per_tok": 8,36 "num_hidden_layers": 4,37 "num_key_value_heads": 4,38 "num_nextn_predict_layers": 1,39 "num_shared_experts": 1,40 "output_router_logits": true,41 "pad_token_id": 120002,42 "qk_norm": true,43 "rms_norm_eps": 1e-05,44 "rope_parameters": {45 "rope_theta": 11158840.0,46 "rope_type": "default"47 },48 "route_norm": true,49 "router_scaling_factor": 2.826,50 "sep_token_id": 120007,51 "tie_word_embeddings": false,52 "transformers_version": "5.13.0",53 "use_cache": true,54 "use_grouped_mm": false,55 "vocab_size": 12083256}57 