Commit
•
8e11fd9
1
Parent(s):
25b5572
Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,27 @@
|
|
| 1 |
---
|
| 2 |
license: llama3
|
| 3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
license: llama3
|
| 3 |
---
|
| 4 |
+
|
| 5 |
+
(Maybe i'll change the waifu picture later)
|
| 6 |
+
|
| 7 |
+
Experimental RP-oriented MoE, the idea was to get a model that would be equal to or better than the Mixtral 8x7B and it's finetunes in RP/ERP tasks.
|
| 8 |
+
|
| 9 |
+
### ChaoticSoliloquy-4x8B
|
| 10 |
+
```
|
| 11 |
+
base_model: jeiku_Chaos_RP_l3_8B
|
| 12 |
+
gate_mode: random
|
| 13 |
+
dtype: bfloat16
|
| 14 |
+
experts_per_token: 2
|
| 15 |
+
experts:
|
| 16 |
+
- source_model: ChaoticNeutrals_Poppy_Porpoise-v0.6-L3-8B
|
| 17 |
+
- source_model: jeiku_Chaos_RP_l3_8B
|
| 18 |
+
- source_model: openlynn_Llama-3-Soliloquy-8B
|
| 19 |
+
- source_model: Sao10K_L3-Solana-8B-v1
|
| 20 |
+
```
|
| 21 |
+
|
| 22 |
+
## Models used
|
| 23 |
+
|
| 24 |
+
- [ChaoticNeutrals/Poppy_Porpoise-v0.6-L3-8B](https://huggingface.co/ChaoticNeutrals/Poppy_Porpoise-v0.6-L3-8B)
|
| 25 |
+
- [jeiku/Chaos_RP_l3_8B](https://huggingface.co/jeiku/Chaos_RP_l3_8B)
|
| 26 |
+
- [openlynn/Llama-3-Soliloquy-8B](https://huggingface.co/openlynn/Llama-3-Soliloquy-8B)
|
| 27 |
+
- [Sao10K/L3-Solana-8B-v1](https://huggingface.co/Sao10K/L3-Solana-8B-v1)
|