LLM EXPLORER 59,420 MODELS INDEXED

Chinese Alpaca 33B Merged by minlik

By minlik · 128 downloads
HF Repo ↗ Copy URL

Chinese Alpaca 33B Merged is an open-source language model by minlik. Features: 33b LLM, VRAM: 65.6GB, Context: 2K, LLM Explorer Score: 0.1, Arc: 59.3, HellaSwag: 78.4, MMLU: 57.7, GSM8K: 8.

  Endpoints compatible   Llama   Pytorch   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/minlik/chinese-alpaca-33b-merged 

Chinese Alpaca 33B Merged Parameters and Internals

Model Type 
text generation, multilingual
Use Cases 
Areas:
research, multilingual text generation
Applications:
text generation tasks involving Chinese language
Primary Use Cases:
multilingual text generation, question answering in Chinese
Limitations:
performance on non-Chinese tasks may vary, accuracy dependent on instruction prompt clarity
Additional Notes 
The model improved Chinese language capabilities by integrating a Chinese vocabulary and fine-tuning with instruction datasets.
Supported Languages 
Chinese (Advanced)
Training Details 
Methodology:
加入中文词表并继续预训练中文Embedding，并在此基础上继续使用指令数据集finetuning
Input Output 
Input Format:
Text-based prompts
Accepted Modalities:
text
Output Format:
Generated text
Performance Tips:
Ensure prompts are clear and precise for best accuracy
LLM NameChinese Alpaca 33B Merged
Repository 🤗https://huggingface.co/minlik/chinese-alpaca-33b-merged 
Model Size33b
Required VRAM65.6 GB
Updated2026-07-22
Maintainerminlik
Model Typellama
Model Files  9.8 GB: 1-of-7   9.9 GB: 2-of-7   10.0 GB: 3-of-7   9.9 GB: 4-of-7   9.9 GB: 5-of-7   9.9 GB: 6-of-7   6.2 GB: 7-of-7   9.8 GB: 1-of-7   9.9 GB: 2-of-7   10.0 GB: 3-of-7   9.9 GB: 4-of-7   9.9 GB: 5-of-7   9.9 GB: 6-of-7   6.2 GB: 7-of-7
Model ArchitectureLlamaForCausalLM
Context Length2048
Model Max Length2048
Transformers Version4.28.1
Tokenizer ClassLlamaTokenizer
Beginning of Sentence Token<s>
End of Sentence Token</s>
Unk Token<unk>
Vocabulary Size49954
Torch Data Typefloat16

Best Alternatives to Chinese Alpaca 33B Merged

Best Alternatives
Context / RAM
Downloads
Likes
Llm Jp 4 33B Thinking64K / 66.8 GB19224
Llm Jp 4 33B Base64K / 66.8 GB898
Llm Jp 4 33B Thinking NVFP464K / 21.6 GB791
Llm Jp 4 33B Thinking Nvfp464K / 21.5 GB382
...angled Llama 33M 32K Base V0.132K / 0.1 GB221
ReflectionCoder DS 33B16K / 67 GB82934
Deepseek Wizard 33B Slerp16K / 35.3 GB70
ValidateAI 33B Slerp16K / 35.4 GB70
ValidateAI 3 33B Ties16K / 66.5 GB70
ValidateAI 2 33B AT16K / 66.5 GB60
Note: green Score (e.g. "73.2") means that the model is better than minlik/chinese-alpaca-33b-merged.