Triangle104 commited on
Commit
16c684f
1 Parent(s): ec19196

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +47 -0
README.md CHANGED
@@ -10,6 +10,53 @@ tags:
10
  This model was converted to GGUF format from [`ArliAI/Mistral-Nemo-12B-ArliAI-RPMax-v1.2`](https://huggingface.co/ArliAI/Mistral-Nemo-12B-ArliAI-RPMax-v1.2) using llama.cpp via the ggml.ai's [GGUF-my-repo](https://huggingface.co/spaces/ggml-org/gguf-my-repo) space.
11
  Refer to the [original model card](https://huggingface.co/ArliAI/Mistral-Nemo-12B-ArliAI-RPMax-v1.2) for more details on the model.
12
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
13
  ## Use with llama.cpp
14
  Install llama.cpp through brew (works on Mac and Linux)
15
 
 
10
  This model was converted to GGUF format from [`ArliAI/Mistral-Nemo-12B-ArliAI-RPMax-v1.2`](https://huggingface.co/ArliAI/Mistral-Nemo-12B-ArliAI-RPMax-v1.2) using llama.cpp via the ggml.ai's [GGUF-my-repo](https://huggingface.co/spaces/ggml-org/gguf-my-repo) space.
11
  Refer to the [original model card](https://huggingface.co/ArliAI/Mistral-Nemo-12B-ArliAI-RPMax-v1.2) for more details on the model.
12
 
13
+ ---
14
+ Model details:
15
+ -
16
+ UPDATE: For those getting gibberish results, it was merged wrongly to base after LORA training. Reuploaded all the files so it should work properly now.
17
+
18
+ RPMax is a series of models that are trained on a diverse set of curated creative writing and RP datasets with a focus on variety and deduplication. This model is designed to be highly creative and non-repetitive by making sure no two entries in the dataset have repeated characters or situations, which makes sure the model does not latch on to a certain personality and be capable of understanding and acting appropriately to any characters or situations.
19
+
20
+ Early tests by users mentioned that these models does not feel like any other RP models, having a different style and generally doesn't feel in-bred.
21
+
22
+ You can access the model at https://arliai.com and ask questions at https://www.reddit.com/r/ArliAI/
23
+
24
+ We also have a models ranking page at https://www.arliai.com/models-ranking
25
+
26
+ Ask questions in our new Discord Server! https://discord.com/invite/t75KbPgwhk
27
+ Model Description
28
+
29
+ ArliAI-RPMax-12B-v1.2 is a variant based on Mistral Nemo 12B Instruct 2407.
30
+
31
+ This is arguably the most successful RPMax model due to how Mistral is already very uncensored in the first place.
32
+
33
+ v1.2 update completely removes non-creative/RP examples in the dataset and is also an incremental improvement of the RPMax dataset which dedups the dataset even more and better filtering to cutout irrelevant description text that came from card sharing sites.
34
+ Specs
35
+
36
+ Context Length: 128K
37
+ Parameters: 12B
38
+
39
+ Training Details
40
+
41
+ Sequence Length: 8192
42
+ Training Duration: Approximately 2 days on 2x3090Ti
43
+ Epochs: 1 epoch training for minimized repetition sickness
44
+ LORA: 64-rank 128-alpha, resulting in ~2% trainable weights
45
+ Learning Rate: 0.00001
46
+ Gradient accumulation: Very low 32 for better learning.
47
+
48
+ Quantization
49
+
50
+ The model is available in quantized formats:
51
+
52
+ FP16: https://huggingface.co/ArliAI/Mistral-Nemo-12B-ArliAI-RPMax-v1.2
53
+ GGUF: https://huggingface.co/ArliAI/Mistral-Nemo-12B-ArliAI-RPMax-v1.2-GGUF
54
+
55
+ Suggested Prompt Format
56
+
57
+ Mistral Instruct Prompt Format
58
+
59
+ ---
60
  ## Use with llama.cpp
61
  Install llama.cpp through brew (works on Mac and Linux)
62