BoreanGale-70B-2.4bpw-h6-exl2

<img src=https://huggingface.co/alchemonaut/BoreanGale-70B/resolve/main/bg.png>

BoreanGale-70B

A merge using a custom algorithm (NearSwap) of:

• 152334H/miqu-1-70b-sf
• Sao10K/WinterGoddess-1.4x-70B-L2

<br/> <br/>

Quants

Several quants are available thanks to community efforts.

NearSwap Algorithm

NearSwap retains most of the weights of the base model (Miqu), but when a weight is similar between the two, it is interpolated to the secondary model (WinterGoddess) value. A parameter t specifies the sameness threshold. When the distance between two values is below t, the weight from the secondary model (WinterGoddess) is used.

This version of the model uses t = 0.001. At this t, about 10% of weights are fully switched to WinterGoddess. Model quality rapidly degrades above t = 0.0025:

• t = 0.0001 (~0.8% full swap): QuartetAnemoi-70B-t0.0001
• t = 0.0003 (~2% full swap)
• t = 0.001 (~10% full swap): This model
• t = 0.0025 (~18% full swap): Generates one paragraph okay, but then reverts to garbage
• t = 0.005 (~35% full swap): Garbage; semi-related word lists
• t = 0.01 (~55% full swap): Garbage; pseudorandom tokens output

NearSwap implementation:

    t: Union[float, np.ndarray],
    v0: Union[np.ndarray, torch.Tensor],
    v1: Union[np.ndarray, torch.Tensor],
...
    lweight = numpy.absolute(v0-v1)
    lweight = t / lweight
    lweight = numpy.nan_to_num(lweight, nan=1.0, posinf=1.0, neginf=1.0)
    numpy.clip(lweight, a_min=0.0, a_max=1.0, out=lweight)
    res = lerp(lweight,v0,v1)

<br/> <br/>

License and Use

Since the ultimate origin of Miqu is at this time unknown beyond speculation, this model is for noncommercial research use only.

<br/> <br/>

Open LLM Leaderboard Evaluation Results

Detailed results can be found here