| { | |
| "_name_or_path": "/home/ubuntu/qiyang/outputs/checkpoint/Falcon3-7B-BitNet-FT-FW-5B/hf_bitnet", | |
| "activation": "swiglu", | |
| "architectures": [ | |
| "LlamaForCausalLM" | |
| ], | |
| "attention_bias": false, | |
| "attention_dropout": 0.0, | |
| "auto_map": { | |
| "AutoConfig": "tiiuae/falcon3-7B--configuration_llama.LlamaConfig", | |
| "AutoModel": "tiiuae/falcon3-7B--modeling_llama.LlamaModel", | |
| "AutoModelForCausalLM": "tiiuae/falcon3-7B--modeling_llama.LlamaForCausalLM", | |
| "AutoModelForQuestionAnswering": "tiiuae/falcon3-7B--modeling_llama.LlamaForQuestionAnswering", | |
| "AutoModelForSequenceClassification": "tiiuae/falcon3-7B--modeling_llama.LlamaForSequenceClassification", | |
| "AutoModelForTokenClassification": "tiiuae/falcon3-7B--modeling_llama.LlamaForTokenClassification" | |
| }, | |
| "bos_token_id": 11, | |
| "eos_token_id": 11, | |
| "head_dim": 256, | |
| "hidden_act": "silu", | |
| "hidden_size": 3072, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 23040, | |
| "is_bitnet_config": true, | |
| "max_position_embeddings": 8192, | |
| "mlp_bias": false, | |
| "model_type": "llama", | |
| "num_attention_heads": 12, | |
| "num_hidden_layers": 28, | |
| "num_key_value_heads": 4, | |
| "parallel_attn": false, | |
| "pretraining_tp": 1, | |
| "quantization_config": { | |
| "modules_to_not_convert": [ | |
| "lm_head" | |
| ], | |
| "quant_method": "bitnet" | |
| }, | |
| "rms_norm_eps": 1e-06, | |
| "rope_scaling": null, | |
| "rope_theta": 1000042, | |
| "tie_word_embeddings": false, | |
| "torch_dtype": "bfloat16", | |
| "transformers_version": "4.46.0.dev0", | |
| "use_cache": true, | |
| "vocab_size": 131072 | |
| } | |