{ "architectures": [ "Qwen2MoeForCausalLM" ], "attention_dropout": 0.0, "bos_token_id": 151643, "decoder_sparse_step": 1, "dtype": "float16", "eos_token_id": 151645, "hidden_act": "silu", "hidden_size": 4, "initializer_range": 0.02, "intermediate_size": 2, "layer_types": [ "sliding_attention", "full_attention" ], "max_position_embeddings": 32768, "max_window_layers": 1, "mlp_only_layers": [], "model_type": "qwen2_moe", "moe_intermediate_size": 2, "norm_topk_prob": false, "num_attention_heads": 4, "num_experts": 60, "num_experts_per_tok": 4, "num_hidden_layers": 2, "num_key_value_heads": 2, "output_router_logits": false, "pad_token_id": null, "qkv_bias": true, "rms_norm_eps": 1e-06, "rope_parameters": { "rope_theta": 1000000.0, "rope_type": "default" }, "router_aux_loss_coef": 0.001, "shared_expert_intermediate_size": 2, "sliding_window": 32768, "tie_word_embeddings": false, "transformers_version": "5.3.0.dev0", "use_cache": true, "use_sliding_window": true, "vocab_size": 151936, "transformers.js_config": { "use_external_data_format": { "model.onnx": 1, "model_fp16.onnx": 1 }, "kv_cache_dtype": { "q4f16": "float16", "fp16": "float16" } } }