Boreas-24B-v1

[!CAUTION] <span style="color:red; font-weight:bold">⚠️ Warning:</span> This merge is partially broken and lobotomized due to bugs with DeepHermesPreview, and is not recommended to use.

[!CAUTION] <span style="color:red; font-weight:bold">⚠️ Note:</span> Use v1.2 specifically instead of v1.0 if you want the same FLUX style. v1.1 and v1.3 use RSCE.

Boreas 24B v1

Initially a test for Goetia, the merge broke due to one bad donor. Possibly there might have also been too many iterations used with the custom flux method which could have overshot the manifold.

As a result, the model has much shorter, summarized outputs than it should have otherwise. With random tool call tags.

It took 61 hours to merge using a 3060 ti. Scores lower than Goetia and not recommended to use, but I wanted to upload the safetensors before they get deleted simply because it took so long.

While this was merging, I actually developed a tool [RSCE Audit] that could check within seconds if models are incompatible or not, thus preventing this waste of time again.

Update The model isn't that bad if you ban the tool call token, maybe has some uses. Consider it another checkpoint model.

Models featured (30)

  - model: A:\LLM\!cache2\BeaverAI_Fallen-Mistral-Small-3.1-24B-v1e_textonly
  - model: A:\LLM\!cache2\Evilmind
  - model: A:\LLM\!cache2\StationV-24B-v1
  - model: A:\LLM\.cache\huggingface\hub\!models--aixonlab--Eurydice-24b-v3.5
  - model: A:\LLM\.cache\huggingface\hub\!models--CrucibleLab--M3.2-24B-Loki-V1.3
  - model: A:\LLM\.cache\huggingface\hub\!models--Darkhn--M3.2-24B-Animus-v7.1
  - model: A:\LLM\.cache\huggingface\hub\!models--Delta-Vector--Austral-24B-Winton
  - model: A:\LLM\.cache\huggingface\hub\!models--Delta-Vector--MS3.2-Austral-Winton
  - model: A:\LLM\.cache\huggingface\hub\!models--Delta-Vector--Rei-24B-KTO
  - model: A:\LLM\.cache\huggingface\hub\!models--Doctor-Shotgun--MS3.2-24B-Magnum-Diamond
  - model: A:\LLM\.cache\huggingface\hub\!models--dphn--Dolphin-Mistral-24B-Venice-Edition
  - model: A:\LLM\.cache\huggingface\hub\!models--FlareRebellion--BereavedCompound-v1.0-24b
  - model: A:\LLM\.cache\huggingface\hub\!models--FlareRebellion--WeirdCompound-v1.7-24b
  - model: A:\LLM\.cache\huggingface\hub\!models--Gryphe--Codex-24B-Small-3.2
  - model: A:\LLM\.cache\huggingface\hub\!models--LatitudeGames--Hearthfire-24B
  - model: A:\LLM\.cache\huggingface\hub\!models--NousResearch--DeepHermes-3-Mistral-24B-Preview
  - model: A:\LLM\.cache\huggingface\hub\!models--OddTheGreat--Circuitry_24B_V.3
  - model: A:\LLM\.cache\huggingface\hub\!models--PocketDoc--Dans-PersonalityEngine-V1.3.0-24b
  - model: A:\LLM\.cache\huggingface\hub\!models--ReadyArt--Broken-Tutu-24B-Transgression-v2.0
  - model: A:\LLM\.cache\huggingface\hub\!models--ReadyArt--MS3.2-The-Omega-Directive-24B-Unslop-v2.1
  - model: A:\LLM\.cache\huggingface\hub\!models--SicariusSicariiStuff--Impish_Magic_24B
  - model: A:\LLM\.cache\huggingface\hub\!models--TheDrummer--Cydonia-24B-v4.3
  - model: A:\LLM\.cache\huggingface\hub\!models--TheDrummer--Magidonia-24B-v4.2.0
  - model: A:\LLM\.cache\huggingface\hub\!models--TheDrummer--Magidonia-24B-v4.3
  - model: A:\LLM\.cache\huggingface\hub\!models--TheDrummer--Precog-24B-v1
  - model: A:\LLM\.cache\huggingface\hub\!models--TheDrummer--Rivermind-24B-v1
  - model: A:\LLM\.cache\huggingface\hub\!models--trashpanda-org--MS3.2-24B-Mullein-v2
  - model: A:\LLM\.cache\huggingface\hub\!models--TroyDoesAI--BlackSheep-24B
  - model: A:\LLM\.cache\huggingface\hub\!models--Undi95--MistralThinker-v1.1
  - model: A:\LLM\.cache\huggingface\hub\!models--zerofata--MS3.2-PaintedFantasy-v3-24B