CyberNative commited on
Commit
184c7b6
·
1 Parent(s): 084d99e

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +7 -0
README.md CHANGED
@@ -7,6 +7,13 @@ license: llama2
7
 
8
  CyberBase 13b 8k *base model* - (llama-2-13b - lmsys/vicuna-13b-v1.5-16k)
9
 
 
 
 
 
 
 
 
10
  # Base cybersecurity model for future fine-tuning, it is not recomended to use on it's own.
11
  - **CyberBase** is a [lmsys/vicuna-13b-v1.5-16k](https://huggingface.co/lmsys/vicuna-13b-v1.5-16k) QLORA fine-tuned on [CyberNative/github_cybersecurity_READMEs](https://huggingface.co/datasets/CyberNative/github_cybersecurity_READMEs) with a single 3090.
12
  - It might, therefore, inherit [promp template of FastChat](https://github.com/lm-sys/FastChat/blob/main/docs/vicuna_weights_version.md#prompt-template)
 
7
 
8
  CyberBase 13b 8k *base model* - (llama-2-13b - lmsys/vicuna-13b-v1.5-16k)
9
 
10
+ ## Test run 1 (less context, more trainable params):
11
+ - sequence_len: 4096
12
+ - max_packed_sequence_len: 4096
13
+ - lora_r: 256
14
+ - lora_alpha: 256
15
+ - trainable params: 1,001,390,080 || all params: 14,017,264,640 || trainable%: 7.143976415643959
16
+
17
  # Base cybersecurity model for future fine-tuning, it is not recomended to use on it's own.
18
  - **CyberBase** is a [lmsys/vicuna-13b-v1.5-16k](https://huggingface.co/lmsys/vicuna-13b-v1.5-16k) QLORA fine-tuned on [CyberNative/github_cybersecurity_READMEs](https://huggingface.co/datasets/CyberNative/github_cybersecurity_READMEs) with a single 3090.
19
  - It might, therefore, inherit [promp template of FastChat](https://github.com/lm-sys/FastChat/blob/main/docs/vicuna_weights_version.md#prompt-template)