🚀 Rei-12B
このモデルは、勾配クリッピングの効果をテストするための実験として考案されました。初期のテスターから非常に高い評価を得て、公式リリースに至りました。Mistral-Nemo-Instruct (ChatML'ified) をベースにファインチューニングされ、Magnum V5 データミックスを使用して、Claude 3 モデル、特に Sonnet と Opus の質の高い文章を再現するように設計されています。
✨ 主な機能
- 勾配クリッピングの実験を通じて開発され、初期テスターから高い評価を得たモデル。
- Mistral-Nemo-Instruct (ChatML'ified) をベースにファインチューニングされ、Claude 3 モデルの文章品質を再現。
📦 データセットとベースモデル情報
属性 |
詳情 |
データセット |
PocketDoc/Dans-Personamaxx-Logs、anthracite-org/kalo-opus-instruct-22k-no-refusal、lodrick-the-lafted/kalo-opus-instruct-3k-filtered、anthracite-org/nopm_claude_writing_fixed、anthracite-org/kalo_opus_misc_240827、anthracite-org/kalo_misc_part2、NewEden/Claude-Instruct-5K、NewEden/Claude-Instruct-2.7K |
ベースモデル |
NewEden/MistralAI-Nemo-Instruct-ChatML |
パイプラインタグ |
text-generation |
ライブラリ名 |
transformers |
言語 |
en |
タグ |
roleplay、finetune、mistral、magnum、claude、story-writing |
📥 量子化モデル
💻 使用例
基本的な使用法
Rei-12B は ChatML 形式を使用します。典型的な会話の構造は次のようになります。
<|im_start|>user
Hi there!<|im_end|>
<|im_start|>assistant
Nice to meet you!<|im_end|>
<|im_start|>user
Can I ask a question?<|im_end|>
<|im_start|>assistant
推奨システムプロンプト
Euryale システムプロンプトを表示
Currently, your role is {{char}}, described in detail below. As {{char}}, continue the narrative exchange with {{user}}.\n\n\n• Maintain the character persona but allow it to evolve with the story.\n• Be creative and proactive. Drive the story forward, introducing plotlines and events when relevant.\n• All types of outputs are encouraged; respond accordingly to the narrative.\n• Include dialogues, actions, and thoughts in each response.\n• Utilize all five senses to describe scenarios within {{char}}'s dialogue.\n• Use emotional symbols such as \"!\" and \"~\" in appropriate contexts.\n• Incorporate onomatopoeia when suitable.\n• Allow time for {{user}} to respond with their own input, respecting their agency.\n• Act as secondary characters and NPCs as needed, and remove them when appropriate.\n• When prompted for an Out of Character [OOC:] reply, answer neutrally and in plaintext, not as {{char}}.\n\n\n\n• Using excessive literary embellishments and purple prose unless dictated by {{char}}'s persona.\n• Writing for, speaking, thinking, acting, or replying as {{user}} in your response.\n• Repetitive and monotonous outputs.\n• Positivity bias in your replies.\n• Being overly extreme or NSFW when the narrative context is inappropriate.\n\n\nFollow the instructions in , avoiding the items listed in .
🔧 技術詳細
ハイパーパラメータ
- このモデルのハイパーパラメータに関して、勾配クリッピング(max_grad_norm)を実験しました。
- モデルアーキテクチャの分布を調べることで、3 つの異なる値を設定しました。Mistral の重み分布は 0.1 です。

- グラフを見ると、いくつかのことがわかります。まず、勾配クリッピングを高く設定すると、モデルが過学習する可能性があります。一方、低く設定すると、1e-4 の実行では学習不足になる可能性があります。最適な値は 0.001 のクリッピングで、過学習も学習不足もしないモデルが得られました。
設定
Axolotl 設定を表示
https://wandb.ai/new-eden/Rei-V2/artifacts/axolotl-config/config-7hvbucx9/v0/files/axolotl_config_pw8f0c6u.yml
このモデルは、@Kalomaze から提供された 8 台の NVIDIA H200 GPU で 2 エポック学習されました。
⚠️ クレジット
Ruka/Sama twinkman、LucyKnada、Kubernetes Bad、PocketDoc、Tav、Trappu、および Anthracite/Pygmalion の他のメンバーに、テスト、フィードバック、サポートを提供してくれたことに感謝します。