This model is on the Llama-3 arch with Llama-3.1 merged in, so it has 8k context length. But could possibly be extended slightly with RoPE due to the L3.1 layers.

Merge Method

This model was merged using the NEARSWAP t0.0001 merge algorithm.

Models Merged

The following models were included in the merge:

Base Model: Casual-Autopsy/L3-Umbral-Mind-RP-v2.0-8B
akjindal53244/Llama-3.1-Storm-8B

Configuration

yaml

slices:
  - sources:
      - model: Casual-Autopsy/L3-Umbral-Mind-RP-v2.0-8B
        layer_range: [0, 32]
      - model: akjindal53244/Llama-3.1-Storm-8B
        layer_range: [0, 32]
merge_method: nearswap
base_model: Casual-Autopsy/L3-Umbral-Mind-RP-v2.0-8B
parameters:
  t:
    - value: 0.0001
dtype: bfloat16

Prompt Template:

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|>

Credit to Alchemonaut:

python

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 to Numbra for idea.