38 lines
930 B
JSON
38 lines
930 B
JSON
{
|
|
"architectures": [
|
|
"GraniteForCausalLM"
|
|
],
|
|
"attention_bias": false,
|
|
"attention_dropout": 0.0,
|
|
"attention_multiplier": 0.0078125,
|
|
"bos_token_id": 49155,
|
|
"dtype": "bfloat16",
|
|
"embedding_multiplier": 12.0,
|
|
"eos_token_id": 0,
|
|
"hidden_act": "silu",
|
|
"hidden_size": 4096,
|
|
"initializer_range": 0.02,
|
|
"intermediate_size": 12800,
|
|
"logits_scaling": 16.0,
|
|
"max_position_embeddings": 131072,
|
|
"mlp_bias": false,
|
|
"model_name": "unsloth/granite-3.2-8b-instruct",
|
|
"model_type": "granite",
|
|
"num_attention_heads": 32,
|
|
"num_hidden_layers": 40,
|
|
"num_key_value_heads": 8,
|
|
"pad_token_id": 4,
|
|
"residual_multiplier": 0.22,
|
|
"rms_norm_eps": 1e-05,
|
|
"rope_parameters": {
|
|
"rope_theta": 10000000.0,
|
|
"rope_type": "default"
|
|
},
|
|
"tie_word_embeddings": true,
|
|
"transformers_version": "5.2.0",
|
|
"unsloth_fixed": true,
|
|
"unsloth_version": "2026.3.3",
|
|
"use_cache": false,
|
|
"vocab_size": 49155
|
|
}
|