Llama 2 70B Longlora 32K by Yukang

 ยป  All LLMs  ยป  Yukang  ยป  Llama 2 70B Longlora 32K   URL Share it on

  Arxiv:2309.12307   Autotrain compatible   Endpoints compatible   Llama   Lora   Region:us

Llama 2 70B Longlora 32K Benchmarks

nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Llama 2 70B Longlora 32K (Yukang/Llama-2-70b-longlora-32k)

Llama 2 70B Longlora 32K Parameters and Internals

Model Type 
Language Model, Text Generation
Additional Notes 
LongLoRA extends models' context while retaining their original architectures, and is compatible with most existing techniques.
Training Details 
Data Sources:
LongQA dataset
Methodology:
Improved fine-tuning using sparse local attention and LoRA for context extension
Hardware Used:
8x A100 machine
Model Architecture:
Retains original architecture, compatible with techniques like FlashAttention-2
Release Notes 
Version:
1.0
Date:
2023
Notes:
Efficient fine-tuning approach using sparse local attention and improved LoRA.
LLM NameLlama 2 70B Longlora 32K
Repository ๐Ÿค—https://huggingface.co/Yukang/Llama-2-70b-longlora-32k 
Model Size70b
Required VRAM0.1 GB
Updated2024-12-23
MaintainerYukang
Model Files  0.1 GB   1.1 GB
Model ArchitectureAutoModelForCausalLM
Is Biasednone
PEFT TypeLORA
LoRA ModelYes
PEFT Target Modulesq_proj|k_proj|v_proj|o_proj
LoRA Alpha16
LoRA Dropout0
R Param8

Best Alternatives to Llama 2 70B Longlora 32K

Best Alternatives
Context / RAM
Downloads
Likes
Llama 3.1 Tango 70B0K / 3 GB867
LLama3 70B SWE LLM0K / 16.1 GB131
Limarpv3 Llama2 70B Qlora0K / 1.7 GB7452
Nous Hermes Llama2 70B0K / 138 GB72883
Llama 2 70B Chat Longlora 32K0K / 0.1 GB139
... 70M Instruct Orca Chkpt 640000K / 0.2 GB151
NorskGPT Llama 3 70B Adapter0K / 0.2 GB95
Llama 3 70B Tagengo0K / 141.9 GB231
...3 70B Instruct Uncensored Lora0K / 0.8 GB243
Note: green Score (e.g. "73.2") means that the model is better than Yukang/Llama-2-70b-longlora-32k.

Rank the Llama 2 70B Longlora 32K 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? 40126 in total.

Our Social Media →  
Original data from HuggingFace, OpenCompass and various public git repos.
Release v20241217