Update config.json to match upstream stablelm-3b-4e1t
#8
by
ndavidson
- opened
- config.json +7 -12
config.json
CHANGED
@@ -1,31 +1,26 @@
|
|
1 |
{
|
2 |
"_name_or_path": "pansophic/rocket-3B",
|
3 |
"architectures": [
|
4 |
-
"
|
5 |
],
|
6 |
-
"auto_map": {
|
7 |
-
"AutoConfig": "stabilityai/stablelm-3b-4e1t--configuration_stablelm_epoch.StableLMEpochConfig",
|
8 |
-
"AutoModelForCausalLM": "stabilityai/stablelm-3b-4e1t--modeling_stablelm_epoch.StableLMEpochForCausalLM"
|
9 |
-
},
|
10 |
"bos_token_id": 0,
|
11 |
-
"eos_token_id":
|
12 |
"hidden_act": "silu",
|
13 |
"hidden_size": 2560,
|
14 |
"initializer_range": 0.02,
|
15 |
"intermediate_size": 6912,
|
16 |
"max_position_embeddings": 4096,
|
17 |
-
"model_type": "
|
18 |
-
"
|
19 |
"num_attention_heads": 32,
|
20 |
-
"num_heads": 32,
|
21 |
"num_hidden_layers": 32,
|
22 |
"num_key_value_heads": 32,
|
23 |
-
"
|
24 |
"rope_theta": 10000,
|
25 |
-
"rotary_scaling_factor": 1.0,
|
26 |
"tie_word_embeddings": false,
|
27 |
"torch_dtype": "bfloat16",
|
28 |
-
"transformers_version": "4.
|
29 |
"use_cache": true,
|
|
|
30 |
"vocab_size": 50304
|
31 |
}
|
|
|
1 |
{
|
2 |
"_name_or_path": "pansophic/rocket-3B",
|
3 |
"architectures": [
|
4 |
+
"StableLmForCausalLM"
|
5 |
],
|
|
|
|
|
|
|
|
|
6 |
"bos_token_id": 0,
|
7 |
+
"eos_token_id": 0,
|
8 |
"hidden_act": "silu",
|
9 |
"hidden_size": 2560,
|
10 |
"initializer_range": 0.02,
|
11 |
"intermediate_size": 6912,
|
12 |
"max_position_embeddings": 4096,
|
13 |
+
"model_type": "stablelm",
|
14 |
+
"layer_norm_eps": 1e-05,
|
15 |
"num_attention_heads": 32,
|
|
|
16 |
"num_hidden_layers": 32,
|
17 |
"num_key_value_heads": 32,
|
18 |
+
"partial_rotary_factor": 0.25,
|
19 |
"rope_theta": 10000,
|
|
|
20 |
"tie_word_embeddings": false,
|
21 |
"torch_dtype": "bfloat16",
|
22 |
+
"transformers_version": "4.38.0",
|
23 |
"use_cache": true,
|
24 |
+
"use_qkv_bias": false,
|
25 |
"vocab_size": 50304
|
26 |
}
|