leviathan-10.5b-final is an open-weight model from Lin, released under Apache License 2.0. It has 4,096-token context. Its published files total 5.9 GB. It draws 21 downloads a month.
這是由 wangguo-2026 開發與封裝的利維坦核心模型 (Leviathan Core)。本模型經過嚴格的監督微調 (SFT)，針對繁體中文邏輯互動與任務執行進行了深度優化。 架構: Llama (Base) 精確度: bfloat16 你可以透過 Hugging Face 函式庫輕鬆載入此模型： 內建 Chat Template，支援標準對話格式。 本模型遵循 Apache 2.0 授權協議。
Model Card
By Lin, published under apache-2.0, revision 95212f7d45d1.
這是由 wangguo-2026 開發與封裝的利維坦核心模型 (Leviathan Core)。本模型經過嚴格的監督微調 (SFT)，針對繁體中文邏輯互動與任務執行進行了深度優化。 架構: Llama (Base) 精確度: bfloat16 你可以透過 Hugging Face 函式庫輕鬆載入此模型： 內建 Chat Template，支援標準對話格式。 本模型遵循 Apache 2.0 授權協議。
Read Lin's full model card
這是由 wangguo-2026 開發與封裝的利維坦核心模型 (Leviathan Core)。本模型經過嚴格的監督微調 (SFT)，針對繁體中文邏輯互動與任務執行進行了深度優化。
模型規格
- 架構: Llama (Base)
- 參數: 10.5B
- 精確度: bfloat16
- 用途: 多任務對話、邏輯推理、繁體中文互動
快速開始
你可以透過 Hugging Face 函式庫輕鬆載入此模型：
from transformers import AutoModelForCausalLM, AutoTokenizer
model_id = "你的使用者名稱/leviathan-10.5b-final"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto")
模型特性
- 內建 Chat Template，支援標準對話格式。
- 針對繁體中文語境進行了對齊，確保語言自然流暢。
授權
本模型遵循 Apache 2.0 授權協議。
Configuration
- Architecture
- LlamaForCausalLM
- Context length (tokens)
- 4,096
- Layers
- 40
- Hidden size
- 4,096
- Feed-forward size
- 11,008
- Attention heads
- 32
- Key/value heads
- 32
- Head dimension
- 128
- Vocabulary size
- 102,400
- Model type
- llama
- Quantization
- bitsandbytes
Identity and Version
- Repository
- Chang-chih/leviathan-10.5b-final
- Publisher
- Lin
- Task
- Not stated by the source
- Modality
- Other
- Library
- Not stated by the source
- Parameters
- Not stated by the source
- Languages
- zh, en
- Revision
- 95212f7d45d1f7b1480b625089ed1530e3a5a332
- First published
- 2026-07-14
- Last updated
- 2026-09-21
Files and Weights
11 files, 5.9 GB in total. The weights are 1 file totalling 5.9 GB in bin.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| pytorch_model.bin | Weights | 5.9 GB | 1638372e18e8 |
| config.json | Configuration | 1.2 KB | — |
| generation_config.json | Configuration | 181 B | — |
| LICENSE | Documentation | 11.4 KB | — |
| README.md | Documentation | 1.0 KB | — |
| Modelfile | Other | 599 B | — |
| chat_template.jinja | Other | 459 B | — |
| 蒸餾的實際流程 | Other | 248 B | — |
| .gitattributes | Repository | 1.5 KB | — |
| tokenizer.json | Tokenizer | 7.5 MB | — |
| tokenizer_config.json | Tokenizer | 396 B | — |
License and Download
- License
- apache-2.0
- Access
- Open weights, no gate
- Download size
- 5.9 GB
Released by Lin through its official repository on Hugging Face. Read the license.
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 5.9 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About leviathan-10.5b-final
Can I use leviathan-10.5b-final commercially?
Yes. leviathan-10.5b-final is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors.
What is leviathan-10.5b-final's context length?
4,096 tokens, from the maximum position embeddings in its published configuration.