| { | |
| "_name_or_path": "unsloth/qwen2.5-3b-instruct-unsloth-bnb-4bit", | |
| "architectures": ["Qwen2ForCausalLM"], | |
| "attention_dropout": 0.0, | |
| "eos_token_id": 151645, | |
| "hidden_act": "silu", | |
| "hidden_size": 2048, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 11008, | |
| "max_position_embeddings": 32768, | |
| "model_type": "qwen2", | |
| "num_attention_heads": 16, | |
| "num_hidden_layers": 36, | |
| "num_key_value_heads": 2, | |
| "pad_token_id": 151654, | |
| "rms_norm_eps": 1e-06, | |
| "rope_theta": 1000000.0, | |
| "torch_dtype": "float16", | |
| "transformers_version": "4.49.0", | |
| "unsloth_fixed": true, | |
| "unsloth_version": "2025.3.14", | |
| "use_cache": true, | |
| "vocab_size": 151936, | |
| "quantization": { | |
| "load_in_4bit": true | |
| }, | |
| "lora": { | |
| "init_lora_weights": true, | |
| "lora_alpha": 64, | |
| "lora_bias": false, | |
| "lora_dropout": 0, | |
| "target_modules": [ | |
| "up_proj", | |
| "o_proj", | |
| "q_proj", | |
| "gate_proj", | |
| "v_proj", | |
| "k_proj", | |
| "down_proj" | |
| ] | |
| } | |
| } | |