trohrbaugh/Stable-DiffCoder-8B-Instruct-heretic

1{2  "architectures": [3    "StableDiffcoderForCausalLM"4  ],5  "attention_bias": false,6  "attention_dropout": 0.1,7  "auto_map": {8    "AutoModel": "modeling_stable_diffcoder.StableDiffcoderForCausalLM",9    "AutoModelForCausalLM": "modeling_stable_diffcoder.StableDiffcoderForCausalLM"10  },11  "bos_token_id": 0,12  "dtype": "bfloat16",13  "eos_token_id": 2,14  "head_dim": 128,15  "hidden_act": "silu",16  "hidden_size": 4096,17  "initializer_range": 0.009882118,18  "intermediate_size": 14336,19  "layer_norm_eps": null,20  "max_position_embeddings": 8192,21  "mlp_bias": false,22  "model_type": "llama",23  "num_attention_heads": 32,24  "num_hidden_layers": 32,25  "num_key_value_heads": 8,26  "pad_token_id": null,27  "pretraining_tp": 1,28  "resid_pdrop": 0.1,29  "rms_norm_eps": 1e-06,30  "rope_parameters": {31    "rope_theta": 500000.0,32    "rope_type": "default"33  },34  "tie_word_embeddings": false,35  "transformers_version": "5.3.0",36  "use_cache": true,37  "vocab_size": 15513638}39