Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,38 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: apache-2.0
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
base_model:
|
| 4 |
+
- allura-org/Teleut-7b
|
| 5 |
+
tags:
|
| 6 |
+
- roleplay
|
| 7 |
+
- conversational
|
| 8 |
+
---
|
| 9 |
+
# Teleut 7b RP
|
| 10 |
+
[cute boygirlthing pending]
|
| 11 |
+
|
| 12 |
+
A roleplay-focused LoRA finetune of Teleut 7b. Methodology and hyperparams inspired by [SorcererLM](https://huggingface.co/rAIfle/SorcererLM-8x22b-bf16).
|
| 13 |
+
|
| 14 |
+
## Dataset
|
| 15 |
+
The worst mix of data you've ever seen. Like, seriously, you do not want to see the things that went into this model. It's bad.
|
| 16 |
+
|
| 17 |
+
## Recommended Settings
|
| 18 |
+
Chat template: ChatML
|
| 19 |
+
Recommended samplers (not the be-all-end-all, try some on your own!):
|
| 20 |
+
- Temp 1.03 / TopK 200 / MinP 0.05 / TopA 0.2
|
| 21 |
+
- Temp 1.03 / TFS 0.75 / TopA 0.3
|
| 22 |
+
|
| 23 |
+
## Hyperparams
|
| 24 |
+
General:
|
| 25 |
+
- Epochs = 2
|
| 26 |
+
- LR = 6e-5
|
| 27 |
+
- LR Scheduler = Cosine
|
| 28 |
+
- Optimizer = Paged AdamW 8bit
|
| 29 |
+
- Effective batch size = 12
|
| 30 |
+
LoRA:
|
| 31 |
+
- Rank = 16
|
| 32 |
+
- Alpha = 32
|
| 33 |
+
- Dropout = 0.25 (Inspiration: [Slush](https://huggingface.co/crestf411/Q2.5-32B-Slush))
|
| 34 |
+
|
| 35 |
+
## Credits
|
| 36 |
+
Thanks to the people who created the data. I would credit you, but that would be cheating ;)
|
| 37 |
+
Thanks to all Allura members, especially Toasty, for testing and emotional support ilya /platonic
|
| 38 |
+
NO thanks to Infermatic. They suck at hosting models
|