58 lines
1.0 KiB
JSON
58 lines
1.0 KiB
JSON
{
|
|
"architectures": [
|
|
"SolarForCausalLM"
|
|
],
|
|
"attention_bias": false,
|
|
"attention_dropout": 0.0,
|
|
"auto_map": {
|
|
"AutoConfig": "configuration_solar.SolarConfig",
|
|
"AutoModelForCausalLM": "modeling_solar.SolarForCausalLM"
|
|
},
|
|
"bos_token_id": 1,
|
|
"bskcn_1": [
|
|
12,
|
|
20,
|
|
32,
|
|
44
|
|
],
|
|
"bskcn_2": [
|
|
20,
|
|
32
|
|
],
|
|
"bskcn_3": [
|
|
16,
|
|
24,
|
|
36,
|
|
48
|
|
],
|
|
"bskcn_4": [
|
|
28,
|
|
40
|
|
],
|
|
"bskcn_tv": [
|
|
0.9,
|
|
0.8
|
|
],
|
|
"eos_token_id": 32007,
|
|
"hidden_act": "silu",
|
|
"hidden_size": 5120,
|
|
"initializer_range": 0.02,
|
|
"intermediate_size": 17920,
|
|
"max_position_embeddings": 4096,
|
|
"mlp_bias": false,
|
|
"model_type": "solar",
|
|
"num_attention_heads": 40,
|
|
"num_hidden_layers": 64,
|
|
"num_key_value_heads": 10,
|
|
"pretraining_tp": 1,
|
|
"rms_norm_eps": 1e-05,
|
|
"rope_scaling": null,
|
|
"rope_theta": 10000.0,
|
|
"sliding_window": 2047,
|
|
"tie_word_embeddings": false,
|
|
"torch_dtype": "bfloat16",
|
|
"transformers_version": "4.44.2",
|
|
"use_cache": true,
|
|
"vocab_size": 32128
|
|
}
|