{ | |
"_name_or_path": "/tmp/models/tmpyhad2wc0/tmpf1vx39wn", | |
"architectures": [ | |
"Transformer" | |
], | |
"auto_map": { | |
"AutoConfig": "LMConfig.LMConfig", | |
"AutoModelForCausalLM": "tclh123/minimind-v1-small--model.Transformer" | |
}, | |
"aux_loss_alpha": 0.01, | |
"dim": 512, | |
"dropout": 0.0, | |
"flash_attn": true, | |
"hidden_dim": null, | |
"max_seq_len": 512, | |
"model_type": "minimind", | |
"multiple_of": 64, | |
"n_heads": 16, | |
"n_kv_heads": 8, | |
"n_layers": 8, | |
"n_routed_experts": 4, | |
"n_shared_experts": true, | |
"norm_eps": 1e-05, | |
"norm_topk_prob": true, | |
"num_experts_per_tok": 2, | |
"quantization_config": { | |
"attention.wk": { | |
"offload_meta": false, | |
"scale_quant_params": null, | |
"weight_quant_params": { | |
"axis": 0, | |
"channel_wise": true, | |
"group_size": 64, | |
"nbits": 8, | |
"optimize": true, | |
"round_zero": false, | |
"view_as_float": false | |
}, | |
"zero_quant_params": { | |
"channel_wise": false, | |
"group_size": null, | |
"nbits": 8, | |
"optimize": false | |
} | |
}, | |
"attention.wo": { | |
"offload_meta": false, | |
"scale_quant_params": null, | |
"weight_quant_params": { | |
"axis": 0, | |
"channel_wise": true, | |
"group_size": 64, | |
"nbits": 8, | |
"optimize": true, | |
"round_zero": false, | |
"view_as_float": false | |
}, | |
"zero_quant_params": { | |
"channel_wise": false, | |
"group_size": null, | |
"nbits": 8, | |
"optimize": false | |
} | |
}, | |
"attention.wq": { | |
"offload_meta": false, | |
"scale_quant_params": null, | |
"weight_quant_params": { | |
"axis": 0, | |
"channel_wise": true, | |
"group_size": 64, | |
"nbits": 8, | |
"optimize": true, | |
"round_zero": false, | |
"view_as_float": false | |
}, | |
"zero_quant_params": { | |
"channel_wise": false, | |
"group_size": null, | |
"nbits": 8, | |
"optimize": false | |
} | |
}, | |
"attention.wv": { | |
"offload_meta": false, | |
"scale_quant_params": null, | |
"weight_quant_params": { | |
"axis": 0, | |
"channel_wise": true, | |
"group_size": 64, | |
"nbits": 8, | |
"optimize": true, | |
"round_zero": false, | |
"view_as_float": false | |
}, | |
"zero_quant_params": { | |
"channel_wise": false, | |
"group_size": null, | |
"nbits": 8, | |
"optimize": false | |
} | |
}, | |
"feed_forward.w1": { | |
"offload_meta": false, | |
"scale_quant_params": null, | |
"weight_quant_params": { | |
"axis": 0, | |
"channel_wise": true, | |
"group_size": 64, | |
"nbits": 8, | |
"optimize": true, | |
"round_zero": false, | |
"view_as_float": false | |
}, | |
"zero_quant_params": { | |
"channel_wise": false, | |
"group_size": null, | |
"nbits": 8, | |
"optimize": false | |
} | |
}, | |
"feed_forward.w2": { | |
"offload_meta": false, | |
"scale_quant_params": null, | |
"weight_quant_params": { | |
"axis": 0, | |
"channel_wise": true, | |
"group_size": 64, | |
"nbits": 8, | |
"optimize": true, | |
"round_zero": false, | |
"view_as_float": false | |
}, | |
"zero_quant_params": { | |
"channel_wise": false, | |
"group_size": null, | |
"nbits": 8, | |
"optimize": false | |
} | |
}, | |
"feed_forward.w3": { | |
"offload_meta": false, | |
"scale_quant_params": null, | |
"weight_quant_params": { | |
"axis": 0, | |
"channel_wise": true, | |
"group_size": 64, | |
"nbits": 8, | |
"optimize": true, | |
"round_zero": false, | |
"view_as_float": false | |
}, | |
"zero_quant_params": { | |
"channel_wise": false, | |
"group_size": null, | |
"nbits": 8, | |
"optimize": false | |
} | |
}, | |
"output": { | |
"offload_meta": false, | |
"scale_quant_params": null, | |
"weight_quant_params": { | |
"axis": 0, | |
"channel_wise": true, | |
"group_size": 64, | |
"nbits": 8, | |
"optimize": true, | |
"round_zero": false, | |
"view_as_float": false | |
}, | |
"zero_quant_params": { | |
"channel_wise": false, | |
"group_size": null, | |
"nbits": 8, | |
"optimize": false | |
} | |
} | |
}, | |
"scoring_func": "softmax", | |
"seq_aux": true, | |
"torch_dtype": "float32", | |
"transformers_version": "4.44.0", | |
"use_moe": false, | |
"vocab_size": 6400 | |
} | |