fwerkor/MiniCPM5-2B-Diffusion-Base
2770
1{2 "architectures": [3 "CIDDiffusionForMaskedLM"4 ],5 "attention_bias": false,6 "attention_dropout": 0.0,7 "base_model_name_or_path": "openbmb/MiniCPM5-2B-Base",8 "bos_token_id": 0,9 "cid_diffusion_base": true,10 "diffusion_attention": "bidirectional",11 "diffusion_objective": "masked-diffusion",12 "dtype": "bfloat16",13 "eos_token_id": 1,14 "head_dim": 128,15 "hidden_act": "silu",16 "hidden_size": 2048,17 "initializer_range": 0.02,18 "intermediate_size": 6144,19 "mask_token_id": 130560,20 "max_position_embeddings": 524288,21 "mlp_bias": false,22 "model_type": "llama",23 "num_attention_heads": 16,24 "num_hidden_layers": 42,25 "num_key_value_heads": 2,26 "pad_token_id": 1,27 "pretraining_tp": 1,28 "rms_norm_eps": 1e-06,29 "rope_scaling": null,30 "rope_theta": 5000000,31 "tie_word_embeddings": false,32 "transformers_version": "4.57.6",33 "use_cache": false,34 "vocab_size": 130561,35 "auto_map": {36 "AutoModelForCausalLM": "modeling_cid_diffusion.CIDDiffusionForMaskedLM"37 }38}39 