| { |
| "_name_or_path": "ServiceNow-AI/Apriel-5B-Instruct", |
| "architectures": [ |
| "AprielForCausalLM" |
| ], |
| "auto_map": { |
| "AutoConfig": "configuration_apriel.AprielConfig", |
| "AutoModelForCausalLM": "modeling_apriel.AprielForCausalLM", |
| "AutoModel": "modeling_apriel.AprielModel" |
| }, |
| "attention_bias": false, |
| "attention_dropout": 0.0, |
| "bos_token_id": 1, |
| "eos_token_id": 2, |
| "head_dim": 128, |
| "hidden_act": "silu", |
| "hidden_size": 4096, |
| "initializer_range": 0.02, |
| "intermediate_size": 8192, |
| "max_position_embeddings": 16384, |
| "mlp_bias": false, |
| "model_type": "apriel", |
| "num_attention_heads": 24, |
| "num_hidden_layers": 28, |
| "num_key_value_heads": 8, |
| "pretraining_tp": 1, |
| "rms_norm_eps": 1e-05, |
| "rope_scaling": { |
| "attention_factor": null, |
| "beta_fast": 32.0, |
| "beta_slow": 1.0, |
| "factor": 32.0, |
| "original_max_position_embeddings": 4096, |
| "rope_type": "yarn" |
| }, |
| "rope_theta": 1000000.0, |
| "tie_word_embeddings": false, |
| "torch_dtype": "bfloat16", |
| "transformers_version": "4.48.3", |
| "use_cache": true, |
| "vocab_size": 131072 |
| } |
|
|