README.md

---
tags:
- llama
- merge
---


Healed Llama-3-15B

![image/png](https://cdn-uploads.huggingface.co/production/uploads/64f74b6e6389380c77562762/Nx5jjEYNH26OS2_87mPTM.png)
# Llama-3-15B-MythicalMaid-t0.0001

## A merge of the following models using a custom NearSwap(t0.0001) algorithm (inverted):

* [ZeusLabs/L3-Aethora-15B-V2](https://huggingface.co/ZeusLabs/L3-Aethora-15B-V2)
* [v000000/HaloMaidRP-v1.33-15B-L3](https://huggingface.co/v000000/HaloMaidRP-v1.33-15B-L3)

With [ZeusLabs/L3-Aethora-15B-V2](https://huggingface.co/ZeusLabs/L3-Aethora-15B-V2) as the base model.

This merge was inverted compared to "L3-15B-EtherealMaid-t0.0001".

## Thanks mradermacher for the quants!
* [GGUF](https://huggingface.co/mradermacher/L3-15B-MythicalMaid-t0.0001-GGUF)
* [GGUF imatrix](https://huggingface.co/mradermacher/L3-15B-MythicalMaid-t0.0001-i1-GGUF)

## Thanks Ransss for the quants!
* [GGUF Q8_0](https://huggingface.co/Ransss/L3-15B-MythicalMaid-t0.0001-Q8_0-GGUF)
* [GGUF Q6_K](https://huggingface.co/Ransss/L3-15B-MythicalMaid-t0.0001-Q6_K-GGUF)

https://huggingface.co/v000000/L3-15B-MythicalMaid-t0.0001-Q5_K_M-GGUF

```python
#Fixed
def lerp(a, b, t):
    return a * (1 - t) + b * t

def nearswap(v0, v1, t):
    lweight = np.abs(v0 - v1)
    with np.errstate(divide='ignore', invalid='ignore'):
        lweight = np.where(lweight != 0, t / lweight, 1.0)
    lweight = np.nan_to_num(lweight, nan=1.0, posinf=1.0, neginf=1.0)
    np.clip(lweight, a_min=0.0, a_max=1.0, out=lweight)
    return lerp(v0, v1, lweight)
```
Credit Alchemonaut

<b>Samplers</b>
```
I found success with:
temperature 0.9-1.2
min_p 0.08
tfs 0.97
smoothing_factor 0.3
smoothing_curve 1.1

Nymeria preset (more coherent):
temp 0.9
top_k 30
top_p 0.75
min_p 0.2
rep_pen 1.1
smooth_factor 0.25
smooth_curve 1

```

<b>Prompt Template</b>
```bash
<|begin_of_text|><|start_header_id|>system<|end_header_id|>

{system_prompt}<|eot_id|><|start_header_id|>user<|end_header_id|>

{input}<|eot_id|><|start_header_id|>assistant<|end_header_id|>

{output}<|eot_id|>

```