RichardErkhov commited on
Commit
4dbb1ae
1 Parent(s): 16643c1

uploaded model

Browse files
Files changed (1) hide show
  1. config.json +50 -0
config.json ADDED
@@ -0,0 +1,50 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "pygmalion-instruct",
3
+ "architectures": [
4
+ "LlamaForCausalLM"
5
+ ],
6
+ "attention_bias": false,
7
+ "attention_dropout": 0.0,
8
+ "badwordsids": [
9
+ [
10
+ 0
11
+ ]
12
+ ],
13
+ "bos_token_id": 1,
14
+ "eos_token_id": 2,
15
+ "hidden_act": "silu",
16
+ "hidden_size": 4096,
17
+ "initializer_range": 0.02,
18
+ "intermediate_size": 11008,
19
+ "max_position_embeddings": 2048,
20
+ "max_sequence_length": 2048,
21
+ "model_type": "llama",
22
+ "num_attention_heads": 32,
23
+ "num_hidden_layers": 32,
24
+ "num_key_value_heads": 32,
25
+ "pad_token_id": 0,
26
+ "pretraining_tp": 1,
27
+ "quantization_config": {
28
+ "_load_in_4bit": false,
29
+ "_load_in_8bit": true,
30
+ "bnb_4bit_compute_dtype": "float32",
31
+ "bnb_4bit_quant_storage": "uint8",
32
+ "bnb_4bit_quant_type": "fp4",
33
+ "bnb_4bit_use_double_quant": false,
34
+ "llm_int8_enable_fp32_cpu_offload": false,
35
+ "llm_int8_has_fp16_weight": false,
36
+ "llm_int8_skip_modules": null,
37
+ "llm_int8_threshold": 6.0,
38
+ "load_in_4bit": false,
39
+ "load_in_8bit": true,
40
+ "quant_method": "bitsandbytes"
41
+ },
42
+ "rms_norm_eps": 1e-06,
43
+ "rope_scaling": null,
44
+ "rope_theta": 10000.0,
45
+ "tie_word_embeddings": false,
46
+ "torch_dtype": "float16",
47
+ "transformers_version": "4.39.3",
48
+ "use_cache": true,
49
+ "vocab_size": 32000
50
+ }