ehartford commited on
Commit
cb79094
·
1 Parent(s): aada896

Upload config.json with huggingface_hub

Browse files
Files changed (1) hide show
  1. config.json +44 -0
config.json ADDED
@@ -0,0 +1,44 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation_function": "gelu_new",
3
+ "architectures": [
4
+ "PhiForCausalLM"
5
+ ],
6
+ "attn_pdrop": 0.0,
7
+ "auto_map": {
8
+ "AutoConfig": "configuration_phi.PhiConfig",
9
+ "AutoModelForCausalLM": "modeling_phi.PhiForCausalLM"
10
+ },
11
+ "embd_pdrop": 0.0,
12
+ "flash_attn": false,
13
+ "flash_rotary": false,
14
+ "fused_dense": false,
15
+ "img_processor": null,
16
+ "initializer_range": 0.02,
17
+ "layer_norm_epsilon": 1e-05,
18
+ "model_type": "phi-msft",
19
+ "n_embd": 2560,
20
+ "n_head": 32,
21
+ "n_head_kv": null,
22
+ "n_inner": null,
23
+ "n_layer": 32,
24
+ "n_positions": 2048,
25
+ "quantization_config": {
26
+ "bnb_4bit_compute_dtype": "float16",
27
+ "bnb_4bit_quant_type": "nf4",
28
+ "bnb_4bit_use_double_quant": true,
29
+ "llm_int8_enable_fp32_cpu_offload": false,
30
+ "llm_int8_has_fp16_weight": false,
31
+ "llm_int8_skip_modules": null,
32
+ "llm_int8_threshold": 6.0,
33
+ "load_in_4bit": true,
34
+ "load_in_8bit": false,
35
+ "quant_method": "bitsandbytes"
36
+ },
37
+ "resid_pdrop": 0.1,
38
+ "rotary_dim": 32,
39
+ "tie_word_embeddings": false,
40
+ "torch_dtype": "float16",
41
+ "transformers_version": "4.37.0.dev0",
42
+ "use_cache": false,
43
+ "vocab_size": 51200
44
+ }