Qwen1.5 110B Chat GPTQ Int4 by study-hjt

 ยป  All LLMs  ยป  study-hjt  ยป  Qwen1.5 110B Chat GPTQ Int4   URL Share it on

  Arxiv:2309.16609   110b   4-bit   4bits   Autotrain compatible Base model:quantized:qwen/qwen... Base model:qwen/qwen1.5-110b-c...   Chat   Conversational   En   Endpoints compatible   Gptq   Int4   Quantized   Qwen   Qwen2   Region:us   Safetensors   Sharded   Tensorflow

Qwen1.5 110B Chat GPTQ Int4 Benchmarks

nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Qwen1.5 110B Chat GPTQ Int4 (study-hjt/Qwen1.5-110B-Chat-GPTQ-Int4)

Qwen1.5 110B Chat GPTQ Int4 Parameters and Internals

Model Type 
text-generation
Additional Notes 
Qwen1.5 is a beta version preceding Qwen2, including enhancements over previous models in performance and multilingual support.
Supported Languages 
en (Full support)
Training Details 
Methodology:
Pretrained with a large dataset; Post-trained with supervised finetuning and direct preference optimization.
Model Architecture:
Transformer-based decoder architecture with SwiGLU activation, attention QKV bias, group query attention, mixture of sliding window attention and full attention. Improved tokenizer adaptive to multiple languages and codes.
Input Output 
Performance Tips:
Use provided hyper-parameters in generation_config.json to avoid code switching or other bad cases.
LLM NameQwen1.5 110B Chat GPTQ Int4
Repository ๐Ÿค—https://huggingface.co/study-hjt/Qwen1.5-110B-Chat-GPTQ-Int4 
Base Model(s)  Qwen/Qwen1.5-110B-Chat   Qwen/Qwen1.5-110B-Chat
Model Size110b
Required VRAM61.5 GB
Updated2025-03-23
Maintainerstudy-hjt
Model Typeqwen2
Model Files  9.8 GB: 1-of-7   9.9 GB: 2-of-7   9.9 GB: 3-of-7   9.9 GB: 4-of-7   9.9 GB: 5-of-7   9.6 GB: 6-of-7   2.5 GB: 7-of-7
Supported Languagesen
GPTQ QuantizationYes
Quantization Typegptq
Model ArchitectureQwen2ForCausalLM
Licenseother
Context Length32768
Model Max Length32768
Transformers Version4.39.3
Tokenizer ClassQwen2Tokenizer
Padding Token<|endoftext|>
Vocabulary Size152064
Torch Data Typefloat16
Errorsreplace

Best Alternatives to Qwen1.5 110B Chat GPTQ Int4

Best Alternatives
Context / RAM
Downloads
Likes
Qwen1.5 110B Chat GPTQ Int432K / 61.7 GB3717
Qwen1.5 110B Chat 4bit32K / 62.2 GB75
Qwen1.5 110B Chat 8bit32K / 179.8 GB51
...n1.5 110B Chat 3.35bpw H6 EXL232K / 49 GB81
...n1.5 110B Chat 3.25bpw H6 EXL232K / 47.7 GB91
Qwen1.5 110B 4bit8K / 62.2 GB41
Qwen1.5 110B Chat32K / 158.3 GB4755125
Qwen1.5 110B32K / 221.7 GB59297
Qwen1.5 110B Chat AWQ32K / 61.7 GB159
Qwen1.5 110B Chat AWQ32K / 61.1 GB80
Note: green Score (e.g. "73.2") means that the model is better than study-hjt/Qwen1.5-110B-Chat-GPTQ-Int4.

Rank the Qwen1.5 110B Chat GPTQ Int4 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? 45433 in total.

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