|
{ |
|
"absolute_position_embedding": false, |
|
"architectures": [ |
|
"MotifForCausalLM" |
|
], |
|
"attention_dropout": 0.0, |
|
"auto_map": { |
|
"AutoConfig": "configuration_motif.MotifConfig", |
|
"AutoModelForCausalLM": "modeling_motif.MotifForCausalLM" |
|
}, |
|
"bos_token_id": 219396, |
|
"eos_token_id": 219395, |
|
"hidden_act": "poly_norm", |
|
"hidden_size": 2048, |
|
"initializer_range": 2e-05, |
|
"intermediate_size": 8192, |
|
"loss_reduction": "mean", |
|
"max_position_embeddings": 16384, |
|
"max_window_layers": 28, |
|
"model_type": "Motif", |
|
"num_attention_heads": 16, |
|
"num_hidden_layers": 32, |
|
"num_key_value_heads": 16, |
|
"rms_norm_eps": 1e-06, |
|
"rope_scaling": null, |
|
"rope_theta": 500000.0, |
|
"sliding_window": null, |
|
"tie_word_embeddings": true, |
|
"torch_dtype": "bfloat16", |
|
"transformers_version": "4.46.3", |
|
"use_bias": false, |
|
"use_cache": true, |
|
"use_sliding_window": false, |
|
"vocab_size": 219520 |
|
} |