LLM EXPLORER
 LLM Explorer: A Curated Large Language Model Directory and Analytics
Dark Theme
  LLM  List    LLM Hosting    LLM Leaderboards    Blog    Newsfeed    Advertise

Azure Dusk V0.2 by Epiculous

 »  All LLMs  »  Epiculous  »  Azure Dusk V0.2   URL Share it on

  36   
  Conversational Dataset:anthracite-org/kalo-op... Dataset:anthracite-org/kalo op... Dataset:anthracite-org/nopm cl... Dataset:anthracite-org/stheno-... Dataset:epiculous/synthrp-gens... Dataset:epiculous/synthstruct-... Dataset:gryphe/sonnet3.5-charc... Dataset:pjmixers/hieunguyenmin...   De   En   Es   Fr   Instruct   It   Ja   Mistral   Model-index   Pt   Region:us   Ru   Safetensors   Sharded   Tensorflow   Zh
Model Card on HF 🤗: https://huggingface.co/Epiculous/Azure_Dusk-v0.2 

Azure Dusk V0.2 Benchmarks

MMLU Pro: 22.6
GPQA: 1.45
MUSR: 6.37
BBH: 17.4
IFEval: 34.67 vs 88 (so35)-60.6%
MATH Lvl 5: 1.81
LLME Score: 0.26871
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Azure Dusk V0.2 (Epiculous/Azure_Dusk-v0.2)

Azure Dusk V0.2 Parameters and Internals

Model Type 
text-generation
Additional Notes 
This version follows Azure_Dusk-v0.2 and includes expanded data and unique training methodology with RSLoRA and ChatML.
Supported Languages 
en (yes), fr (yes), de (yes), es (yes), it (yes), pt (yes), ru (yes), zh (yes), ja (yes)
Training Details 
Data Sources:
Epiculous/SynthRP-Gens-v1.1-Filtered-n-Cleaned, anthracite-org/stheno-filtered-v1.1, PJMixers/hieunguyenminh_roleplay-deduped-ShareGPT, Gryphe/Sonnet3.5-Charcard-Roleplay, Epiculous/Synthstruct-Gens-v1.1-Filtered-n-Cleaned, anthracite-org/kalo-opus-instruct-22k-no-refusal, anthracite-org/nopm_claude_writing_fixed, anthracite-org/kalo_opus_misc_240827
Methodology:
Training was done using RSLoRA instead of regular LoRA. A two-phased approach was utilized; first training on RP data followed by instruct data.
Training Time:
Twice over 2 epochs each on two 2x NVIDIA A6000 GPUs
Hardware Used:
2x NVIDIA A6000 GPUs
Model Architecture:
ChatML formatting with Changed prompting structure.
Input Output 
Input Format:
Trained on ChatML formatting.
Accepted Modalities:
text
Output Format:
text
LLM NameAzure Dusk V0.2
Repository 🤗https://huggingface.co/Epiculous/Azure_Dusk-v0.2 
Model Size12.2b
Required VRAM24.5 GB
Updated2024-12-30
MaintainerEpiculous
Model Typemistral
Instruction-BasedYes
Model Files  4.9 GB: 1-of-5   4.9 GB: 2-of-5   4.9 GB: 3-of-5   4.9 GB: 4-of-5   4.9 GB: 5-of-5
Supported Languagesen fr de es it pt ru zh ja
Model ArchitectureMistralForCausalLM
Licenseapache-2.0
Context Length1024000
Model Max Length1024000
Transformers Version4.44.0
Tokenizer ClassPreTrainedTokenizerFast
Padding Token<pad>
Vocabulary Size131072
Torch Data Typebfloat16

Best Alternatives to Azure Dusk V0.2

Best Alternatives
Context / RAM
Downloads
Likes
Violet Twilight V0.21000K / 24.5 GB119921
...ish Mistral Nemo Instruct 24071000K / 24.5 GB772
Mistral Nemo Kurdish1000K / 24.5 GB442
...al Nemo Japanese Instruct 24081000K / 24.5 GB208427
Mistral Nemo Instruct 24071000K / 24.5 GB90746
...l Nemo Abliterated Nemo Pro V21000K / 24.5 GB2350
Crimson Dawn V0.21000K / 24.5 GB5110
MN SlushoMix1000K / 24.5 GB332
ChatML Nemo Pro1000K / 24.5 GB281
10 14dpo1000K / 49.3 GB1621
Note: green Score (e.g. "73.2") means that the model is better than Epiculous/Azure_Dusk-v0.2.

Rank the Azure Dusk V0.2 Capabilities

🆘 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! 🌟

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? 40476 in total.

Email us: info@extractum.io. Our Privacy Policy  | Terms and Conditions  | Suggest an improvement.
Our Social Media →  
Original data from HuggingFace, OpenCompass and various public git repos.
Release v20241227
  Support LLM Explorer