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

KoSoLAR 10.7B V0.2 1.4 Dedup by jingyeom

 ยป  All LLMs  ยป  jingyeom  ยป  KoSoLAR 10.7B V0.2 1.4 Dedup   URL Share it on

  68   
  32bit   Autotrain compatible   Conversational   Endpoints compatible   Llama   Quantized   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF ๐Ÿค—: https://huggingface.co/jingyeom/KoSoLAR-10.7B-v0.2_1.4_dedup 

KoSoLAR 10.7B V0.2 1.4 Dedup Benchmarks

ARC: 63.05 vs 96.7 (so35)-34.8%
HellaSwag: 83.63 vs 95.3 (gpt4)-12.2%
MMLU: 64.61 vs 88.3 (so35)-26.8%
TruthfulQA: 52.69 vs 59 (gpt4)-10.7%
WinoGrande: 80.51 vs 87.5 (gpt4)-8%
GSM8K: 48.07 vs 96.4 (so35)-50.1%
LLME Score: 0.18426
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
KoSoLAR 10.7B V0.2 1.4 Dedup (jingyeom/KoSoLAR-10.7B-v0.2_1.4_dedup)

KoSoLAR 10.7B V0.2 1.4 Dedup Parameters and Internals

Training Details 
Data Sources:
๊ณต๊ฐœ ๋ฐ์ดํ„ฐ ์ˆ˜์ง‘, Deduplicating Training Data Makes Language Models Better ์•Œ๊ณ ๋ฆฌ์ฆ˜ ํ™œ์šฉ
LLM NameKoSoLAR 10.7B V0.2 1.4 Dedup
Repository ๐Ÿค—https://huggingface.co/jingyeom/KoSoLAR-10.7B-v0.2_1.4_dedup 
Base Model(s)  jingyeom/KoSoLAR-10.7B-v0.2_1.3_dedup_p   jingyeom/KoSoLAR-10.7B-v0.2_1.3_dedup_p
Model Size10.7b
Required VRAM21.6 GB
Updated2025-01-15
Maintainerjingyeom
Model Typellama
Model Files  4.9 GB: 1-of-5   4.9 GB: 2-of-5   5.0 GB: 3-of-5   4.9 GB: 4-of-5   1.9 GB: 5-of-5
Quantization Type32bit
Model ArchitectureLlamaForCausalLM
Licenseapache-2.0
Context Length4096
Model Max Length4096
Transformers Version4.35.2
Tokenizer ClassLlamaTokenizer
Padding Token</s>
Vocabulary Size40960
Torch Data Typefloat16

Best Alternatives to KoSoLAR 10.7B V0.2 1.4 Dedup

Best Alternatives
Context / RAM
Downloads
Likes
... 10.7B Instruct V1.0 128K EXL2128K / 5.6 GB181
KoSoLAR 10.7B V0.2 1.4 Dedup 14K / 21.6 GB33230
...e KoSoLAR 10.7B V0.2 1.4 Dedup4K / 43.2 GB23240
Solar Test 240517 16bit4K / 21.4 GB2610
Eclectus 1.1 Dedup4K / 21.4 GB33070
Eclectus1.14K / 21.4 GB33150
Frostwind 10.7B V1 EXL24K / 10.9 GB121
Solar 10.7B Bnb 4bit4K / 5.9 GB703
Kiwi Solar Merge Ties2 DPO4K / 21.4 GB1610
...ess 10.7B V1.5B 8.0bpw H8 EXL24K / 10.9 GB91
Note: green Score (e.g. "73.2") means that the model is better than jingyeom/KoSoLAR-10.7B-v0.2_1.4_dedup.

Rank the KoSoLAR 10.7B V0.2 1.4 Dedup 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? 41363 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