Airoboros DPO 110B 3.3 by jondurbin

 ยป  All LLMs  ยป  jondurbin  ยป  Airoboros DPO 110B 3.3   URL Share it on

  Autotrain compatible   Conversational Dataset:allenai/ultrafeedback ... Dataset:argilla/distilabel-int... Dataset:bluemoon-fandom-1-1-rp...   Dataset:boolq Dataset:glaiveai/glaive-functi... Dataset:grimulkan/limarp-augme... Dataset:jondurbin/airoboros-3.... Dataset:jondurbin/cinematika-v... Dataset:jondurbin/contextual-d... Dataset:jondurbin/gutenberg-dp...   Dataset:jondurbin/py-dpo-v0.1 Dataset:jondurbin/truthy-dpo-v...   Dataset:ldjnr/capybara   Dataset:lmsys/lmsys-chat-1m Dataset:mattpscott/airoboros-s...   Dataset:piqa Dataset:unalignment/toxic-dpo-... Dataset:vezora/tested-22k-pyth...   Endpoints compatible   Qwen2   Region:us   Safetensors   Sharded   Tensorflow

Airoboros DPO 110B 3.3 Benchmarks

nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Airoboros DPO 110B 3.3 (jondurbin/airoboros-dpo-110b-3.3)

Airoboros DPO 110B 3.3 Parameters and Internals

Model Type 
text generation
Use Cases 
Areas:
research
Applications:
text generation, coding, summarization, agent/function calling, contextual question answering
Primary Use Cases:
Question answering, Coding assistance, Summarization, Agent/function calling
Limitations:
The model sometimes adds random extra tokens at the end of responses. It uses a closed-context formatting for more accurate outputs.
Considerations:
It is strongly advised to avoid commercial usage due to the OpenAI API usage.
Additional Notes 
This model is experimental and its responses include some token randomness, requiring special prompt formatting for closed-context tasks.
Training Details 
Data Sources:
jondurbin/airoboros-3.2, bluemoon-fandom-1-1-rp-cleaned, boolq, LDJnr/Capybara, jondurbin/cinematika-v0.1, glaiveai/glaive-function-calling-v2, grimulkan/LimaRP-augmented, piqa, Vezora/Tested-22k-Python-Alpaca, mattpscott/airoboros-summarization, unalignment/toxic-dpo-v0.2, allenai/ultrafeedback_binarized_cleaned, argilla/distilabel-intel-orca-dpo-pairs, jondurbin/contextual-dpo-v0.1, jondurbin/gutenberg-dpo-v0.1, jondurbin/py-dpo-v0.1, jondurbin/truthy-dpo-v0.1, lmsys/lmsys-chat-1m
Methodology:
Fine-tuning through ChatML prompt formatting using DPO pass and synthetic data generated by Airoboros.
Input Output 
Input Format:
ChatML prompt template, using 'apply_chat_template' method for accuracy.
Accepted Modalities:
text
Output Format:
Text with structured response format using JSON/YAML for specific tasks.
Performance Tips:
Use a low temperature setting for contextually obedient responses.
LLM NameAiroboros DPO 110B 3.3
Repository ๐Ÿค—https://huggingface.co/jondurbin/airoboros-dpo-110b-3.3 
Model Size110b
Required VRAM158.3 GB
Updated2025-03-15
Maintainerjondurbin
Model Typeqwen2
Model Files  3.6 GB: 1-of-62   3.5 GB: 2-of-62   3.8 GB: 3-of-62   3.5 GB: 4-of-62   3.5 GB: 5-of-62   3.8 GB: 6-of-62   3.5 GB: 7-of-62   3.5 GB: 8-of-62   3.8 GB: 9-of-62   3.5 GB: 10-of-62   3.5 GB: 11-of-62   3.8 GB: 12-of-62   3.5 GB: 13-of-62   3.5 GB: 14-of-62   3.8 GB: 15-of-62   3.5 GB: 16-of-62   3.5 GB: 17-of-62   3.8 GB: 18-of-62   3.5 GB: 19-of-62   3.5 GB: 20-of-62   3.8 GB: 21-of-62   3.5 GB: 22-of-62   3.5 GB: 23-of-62   3.8 GB: 24-of-62   3.5 GB: 25-of-62   3.5 GB: 26-of-62   3.8 GB: 27-of-62   3.5 GB: 28-of-62   3.5 GB: 29-of-62   3.8 GB: 30-of-62   3.5 GB: 31-of-62   3.5 GB: 32-of-62   3.8 GB: 33-of-62   3.5 GB: 34-of-62   3.5 GB: 35-of-62   3.8 GB: 36-of-62   3.5 GB: 37-of-62   3.5 GB: 38-of-62   3.8 GB: 39-of-62   3.5 GB: 40-of-62   3.5 GB: 41-of-62   3.8 GB: 42-of-62   3.5 GB: 43-of-62   3.5 GB: 44-of-62
Model ArchitectureQwen2ForCausalLM
Licenseother
Context Length32768
Model Max Length32768
Transformers Version4.40.1
Vocabulary Size152064
Torch Data Typebfloat16

Best Alternatives to Airoboros DPO 110B 3.3

Best Alternatives
Context / RAM
Downloads
Likes
Qwen1.5 110B Chat32K / 158.3 GB4880125
Qwen1.5 110B32K / 221.7 GB68797
Dolphin 2.9.1 Qwen 110B32K / 193.4 GB4228
Airoboros 110B 3.332K / 158.3 GB102
Aqua Qwen 0.1 110B32K / 198.3 GB60
...wen 1.5 110B Layer Mix Bpw 2.28K / 40.7 GB91
Qwen1.5 110B Chat 4bit32K / 62.2 GB85
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
Note: green Score (e.g. "73.2") means that the model is better than jondurbin/airoboros-dpo-110b-3.3.

Rank the Airoboros DPO 110B 3.3 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? 45097 in total.

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