PocketDoc commited on
Commit
108efb1
1 Parent(s): e66ecba

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +4 -4
README.md CHANGED
@@ -4,14 +4,14 @@ license: llama2
4
  # LLaMA2-13B-TiefighterLR
5
  TiefighterLR is a merged model achieved trough merging two different lora's on top of a well established existing merge.
6
  This LR version contains Less Rodeo, merged at 3% from the original 5% reducing its second person adventure bias.
7
- Testers found this model to understand your own character and instruction prompts better, at the sacrefise of lowering its own writing bias/style.
8
 
9
  To achieve this the following recipe was used:
10
 
11
  * We begin with the base model Undi95/Xwin-MLewd-13B-V0.2 which is a well established merge, contrary to the name this model does not have a strong NSFW bias.
12
  * Then we applied the PocketDoc/Dans-RetroRodeo-13b lora which is a finetune on the Choose your own Adventure datasets from our Skein model.
13
  * After applying this lora we merged the original model with the newly created PocketDoc/Dans-RetroRodeo-13b merge at 3% to weaken the newly introduced adventure bias.
14
- * The resulting merge was used as a new basemodel to which we applied Blackroot/Llama-2-13B-Storywriter-LORA and repeated the same trick, this time at 10%.
15
 
16
  This means this model contains the following ingredients from their upstream models for as far as we can track them:
17
  - Undi95/Xwin-MLewd-13B-V0.2
@@ -50,14 +50,14 @@ Regular story writing in the traditional way is supported, simply copy paste you
50
  To generate stories on demand you can use an instruction (tested in the Alpaca format) such as "Write a novel about X, use chapters and dialogue" this will generate a story. The format can vary between generations depending on how the model chooses to begin, either write what you want as shown in the earlier example or write the beginning of the story yourself so the model can follow your style. A few retries can also help if the model gets it wrong.
51
 
52
  ## Chatbots and persona's
53
- Unlike the original Tiefigther this model is better at handling excisting Character Cards as long as they do not contain a lot of second person writing or second person introductions (You), setting > as a custom stop sequence can help fix potential mistakes, as well as turning multi-line replies off.
54
  You can also use instructions to create your characters.
55
 
56
  For example, you can put this in memory in regular chat mode:
57
  ```
58
  ### Instruction:
59
  Generate a conversation between Alice and Henk where they discuss language models.
60
- In this conversation Henk is excited to teach Alice about Tiefigther.
61
  ### Response:
62
  ```
63
 
 
4
  # LLaMA2-13B-TiefighterLR
5
  TiefighterLR is a merged model achieved trough merging two different lora's on top of a well established existing merge.
6
  This LR version contains Less Rodeo, merged at 3% from the original 5% reducing its second person adventure bias.
7
+ Testers found this model to understand your own character and instruction prompts better, at the sacrifice of lowering its own writing bias/style.
8
 
9
  To achieve this the following recipe was used:
10
 
11
  * We begin with the base model Undi95/Xwin-MLewd-13B-V0.2 which is a well established merge, contrary to the name this model does not have a strong NSFW bias.
12
  * Then we applied the PocketDoc/Dans-RetroRodeo-13b lora which is a finetune on the Choose your own Adventure datasets from our Skein model.
13
  * After applying this lora we merged the original model with the newly created PocketDoc/Dans-RetroRodeo-13b merge at 3% to weaken the newly introduced adventure bias.
14
+ * The resulting merge was used as a new base model to which we applied Blackroot/Llama-2-13B-Storywriter-LORA and repeated the same trick, this time at 10%.
15
 
16
  This means this model contains the following ingredients from their upstream models for as far as we can track them:
17
  - Undi95/Xwin-MLewd-13B-V0.2
 
50
  To generate stories on demand you can use an instruction (tested in the Alpaca format) such as "Write a novel about X, use chapters and dialogue" this will generate a story. The format can vary between generations depending on how the model chooses to begin, either write what you want as shown in the earlier example or write the beginning of the story yourself so the model can follow your style. A few retries can also help if the model gets it wrong.
51
 
52
  ## Chatbots and persona's
53
+ Unlike the original Tiefighter this model is better at handling existing Character Cards as long as they do not contain a lot of second person writing or second person introductions (You), setting > as a custom stop sequence can help fix potential mistakes, as well as turning multi-line replies off.
54
  You can also use instructions to create your characters.
55
 
56
  For example, you can put this in memory in regular chat mode:
57
  ```
58
  ### Instruction:
59
  Generate a conversation between Alice and Henk where they discuss language models.
60
+ In this conversation Henk is excited to teach Alice about Tiefighter.
61
  ### Response:
62
  ```
63