Upload config.json with huggingface_hub
Browse files- config.json +14 -9
config.json
CHANGED
|
@@ -1,26 +1,31 @@
|
|
| 1 |
{
|
| 2 |
"architectures": [
|
| 3 |
-
"
|
| 4 |
],
|
|
|
|
|
|
|
| 5 |
"auto_map": {
|
| 6 |
"AutoConfig": "configuration_aquila.AquilaConfig",
|
| 7 |
"AutoModelForCausalLM": "modeling_aquila.AquilaForCausalLM"
|
| 8 |
},
|
| 9 |
-
"bos_token_id":
|
| 10 |
-
"eos_token_id":
|
| 11 |
"hidden_act": "silu",
|
| 12 |
"hidden_size": 4096,
|
| 13 |
"initializer_range": 0.02,
|
| 14 |
"intermediate_size": 11008,
|
| 15 |
-
"max_position_embeddings":
|
| 16 |
"model_type": "aquila",
|
| 17 |
"num_attention_heads": 32,
|
| 18 |
"num_hidden_layers": 32,
|
| 19 |
-
"
|
|
|
|
| 20 |
"rms_norm_eps": 1e-05,
|
|
|
|
|
|
|
| 21 |
"tie_word_embeddings": false,
|
| 22 |
-
"torch_dtype": "
|
| 23 |
-
"transformers_version": "4.
|
| 24 |
"use_cache": true,
|
| 25 |
-
"vocab_size":
|
| 26 |
-
}
|
|
|
|
| 1 |
{
|
| 2 |
"architectures": [
|
| 3 |
+
"AquilaForCausalLM"
|
| 4 |
],
|
| 5 |
+
"attention_bias": false,
|
| 6 |
+
"attention_dropout": 0.0,
|
| 7 |
"auto_map": {
|
| 8 |
"AutoConfig": "configuration_aquila.AquilaConfig",
|
| 9 |
"AutoModelForCausalLM": "modeling_aquila.AquilaForCausalLM"
|
| 10 |
},
|
| 11 |
+
"bos_token_id": 143717,
|
| 12 |
+
"eos_token_id": 143718,
|
| 13 |
"hidden_act": "silu",
|
| 14 |
"hidden_size": 4096,
|
| 15 |
"initializer_range": 0.02,
|
| 16 |
"intermediate_size": 11008,
|
| 17 |
+
"max_position_embeddings": 8192,
|
| 18 |
"model_type": "aquila",
|
| 19 |
"num_attention_heads": 32,
|
| 20 |
"num_hidden_layers": 32,
|
| 21 |
+
"num_key_value_heads": 32,
|
| 22 |
+
"pretraining_tp": 1,
|
| 23 |
"rms_norm_eps": 1e-05,
|
| 24 |
+
"rope_scaling": null,
|
| 25 |
+
"rope_theta": 1000000.0,
|
| 26 |
"tie_word_embeddings": false,
|
| 27 |
+
"torch_dtype": "bfloat16",
|
| 28 |
+
"transformers_version": "4.39.3",
|
| 29 |
"use_cache": true,
|
| 30 |
+
"vocab_size": 143973
|
| 31 |
+
}
|