RichardErkhov commited on
Commit
dbbbd08
1 Parent(s): e0698be

uploaded model

Browse files
Files changed (1) hide show
  1. config.json +39 -0
config.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "granite-3b-code-base-128k",
3
+ "architectures": [
4
+ "LlamaForCausalLM"
5
+ ],
6
+ "attention_bias": true,
7
+ "attention_dropout": 0.1,
8
+ "bos_token_id": 0,
9
+ "eos_token_id": 0,
10
+ "head_dim": 80,
11
+ "hidden_act": "silu",
12
+ "hidden_size": 2560,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 10240,
15
+ "max_position_embeddings": 128000,
16
+ "mlp_bias": true,
17
+ "model_type": "llama",
18
+ "num_attention_heads": 32,
19
+ "num_hidden_layers": 32,
20
+ "num_key_value_heads": 32,
21
+ "pad_token_id": 0,
22
+ "pretraining_tp": 1,
23
+ "quantization_config": {
24
+ "bits": 4,
25
+ "group_size": 128,
26
+ "modules_to_not_convert": null,
27
+ "quant_method": "awq",
28
+ "version": "gemm",
29
+ "zero_point": true
30
+ },
31
+ "rms_norm_eps": 1e-05,
32
+ "rope_scaling": null,
33
+ "rope_theta": 10000000,
34
+ "tie_word_embeddings": true,
35
+ "torch_dtype": "float16",
36
+ "transformers_version": "4.46.3",
37
+ "use_cache": false,
38
+ "vocab_size": 49152
39
+ }