Qwen2.5 14B Bnb 4bit by unsloth

 ยป  All LLMs  ยป  unsloth  ยป  Qwen2.5 14B Bnb 4bit   URL Share it on

  Arxiv:2407.10671   4-bit   4bit   Autotrain compatible Base model:quantized:qwen/qwen...   Base model:qwen/qwen2.5-14b   Bitsandbytes   En   Endpoints compatible   Quantized   Qwen2   Region:us   Safetensors   Sharded   Tensorflow   Unsloth

Qwen2.5 14B Bnb 4bit Benchmarks

nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Qwen2.5 14B Bnb 4bit (unsloth/Qwen2.5-14B-bnb-4bit)

Qwen2.5 14B Bnb 4bit Parameters and Internals

Model Type 
Causal Language Model
Use Cases 
Areas:
Research, Coding, Mathematics, Multilingual Applications
Applications:
Chatbots, Structured Data Understanding, Role-playing Implementations
Primary Use Cases:
instruction following, generating long texts, understanding structured data, generating structured outputs
Limitations:
Not recommended for conversations without further training
Considerations:
Use post-training methods for conversations like SFT, RLHF
Additional Notes 
Supports multilingual capabilities with improved role-play conditions for chatbots.
Supported Languages 
English (high), Chinese (high), French (medium), Spanish (medium), Portuguese (medium), German (medium), Italian (medium), Russian (medium), Japanese (medium), Korean (medium), Vietnamese (medium), Thai (medium), Arabic (medium), other_languages (basic)
Training Details 
Data Sources:
specialized expert models in coding and mathematics
Context Length:
131072
Model Architecture:
transformers with RoPE, SwiGLU, RMSNorm, and Attention QKV bias
Input Output 
Input Format:
text
Accepted Modalities:
text
Output Format:
text
Performance Tips:
Use the latest version of 'transformers' to avoid errors.
LLM NameQwen2.5 14B Bnb 4bit
Repository ๐Ÿค—https://huggingface.co/unsloth/Qwen2.5-14B-bnb-4bit 
Base Model(s)  Qwen/Qwen2.5-14B   Qwen/Qwen2.5-14B
Model Size14b
Required VRAM9.9 GB
Updated2025-02-10
Maintainerunsloth
Model Typeqwen2
Model Files  5.0 GB: 1-of-2   4.9 GB: 2-of-2
Supported Languagesen
Quantization Type4bit
Model ArchitectureQwen2ForCausalLM
Licenseapache-2.0
Context Length131072
Model Max Length131072
Transformers Version4.44.2
Tokenizer ClassQwen2Tokenizer
Padding Token<|PAD_TOKEN|>
Vocabulary Size152064
Torch Data Typebfloat16
Errorsreplace

Best Alternatives to Qwen2.5 14B Bnb 4bit

Best Alternatives
Context / RAM
Downloads
Likes
...B Instruct 1M Unsloth Bnb 4bit986K / 14.3 GB3722
Qwen2.5 14B Instruct 1M 8bit986K / 15.7 GB3422
...till Qwen 14B Unsloth Bnb 4bit128K / 13.9 GB1518715
QwQ 14B Math V0.2128K / 29.7 GB14715
...k R1 Distill Qwen 14B Bnb 4bit128K / 9.9 GB38993
Qwen Carpmuscle V0.3128K / 29.7 GB50
...1 Distill Qwen 14B EXL2 8.0bpw128K / 14.7 GB2162
UwU 14B Math V0.2128K / 29.7 GB19818
...antha2.0 Qwen2.5 14B Ita 16bit128K / 29.7 GB27360
...ha2.0 Qwen2.5 14B Ita V2 16bit128K / 29.7 GB25150
Note: green Score (e.g. "73.2") means that the model is better than unsloth/Qwen2.5-14B-bnb-4bit.

Rank the Qwen2.5 14B Bnb 4bit Capabilities

๐Ÿ†˜ Have you tried this model? Rate its performance. This feedback would greatly assist ML community in identifying the most suitable model for their needs. Your contribution really does make a difference! ๐ŸŒŸ

Instruction Following and Task Automation  
Factuality and Completeness of Knowledge  
Censorship and Alignment  
Data Analysis and Insight Generation  
Text Generation  
Text Summarization and Feature Extraction  
Code Generation  
Multi-Language Support and Translation  

What open-source LLMs or SLMs are you in search of? 42935 in total.

Our Social Media →  
Original data from HuggingFace, OpenCompass and various public git repos.
Release v20241227