Llama 33B Lxctx PI 16384 LoRA by bhenrym14

 ยป  All LLMs  ยป  bhenrym14  ยป  Llama 33B Lxctx PI 16384 LoRA   URL Share it on

  Adapter   Finetuned   Lora   Region:us

Llama 33B Lxctx PI 16384 LoRA Benchmarks

nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Llama 33B Lxctx PI 16384 LoRA (bhenrym14/llama-33b-lxctx-PI-16384-LoRA)
๐ŸŒŸ Advertise your project ๐Ÿš€

Llama 33B Lxctx PI 16384 LoRA Parameters and Internals

Model Type 
text generation
Additional Notes 
This is a QLoRA fine-tune focusing on extending the context window to 16384 by using RoPE Scaled Embeddings for Position Interpolation. Pretrained for additional 100 steps on 8192 length sequences.
Training Details 
Data Sources:
the pile dataset
Methodology:
RoPE Scaled Embeddings (Position Interpolation)
Context Length:
16384
Training Time:
10 hours
Hardware Used:
1x RTX 6000 Ada
LLM NameLlama 33B Lxctx PI 16384 LoRA
Repository ๐Ÿค—https://huggingface.co/bhenrym14/llama-33b-lxctx-PI-16384-LoRA 
Model Size33b
Required VRAM1 GB
Updated2025-09-14
Maintainerbhenrym14
Model Files  1.0 GB
Model ArchitectureAdapter
Is Biasednone
PEFT TypeLORA
LoRA ModelYes
PEFT Target Moduleso_proj|k_proj|q_proj|up_proj|down_proj|v_proj|gate_proj
LoRA Alpha16
LoRA Dropout0.05
R Param64

Best Alternatives to Llama 33B Lxctx PI 16384 LoRA

Best Alternatives
Context / RAM
Downloads
Likes
MentaLLaMA 33B Lora0K / 0.1 GB05
Airoboros 33B 2.1 Peft0K / 1 GB01
Chinese Alpaca Pro Lora 33B0K / 2.3 GB07
Chinese Alpaca Plus Lora 33B0K / 2.3 GB02
Sorceroboros 33B S2a4 Qlora0K / 1.9 GB11
...Gpt4 1.4.1 Lxctx PI 16384 LoRA0K / 1 GB01
Guanaco Unchained 33B Qlora0K / 1 GB03
Enterredaas 33B QLoRA0K / 1 GB04
... 33B Gpt4 1.4.1 NTK 16384 LoRA0K / 1 GB02
Chinese Alpaca Lora 33B0K / 2.9 GB010
Note: green Score (e.g. "73.2") means that the model is better than bhenrym14/llama-33b-lxctx-PI-16384-LoRA.

Rank the Llama 33B Lxctx PI 16384 LoRA 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? 51368 in total.

Our Social Media →  
Original data from HuggingFace, OpenCompass and various public git repos.
Release v20241124