Phi 2 GPTQ by TheBloke

 ยป  All LLMs  ยป  TheBloke  ยป  Phi 2 GPTQ   URL Share it on

  4-bit   Autotrain compatible   Base model:microsoft/phi-2 Base model:quantized:microsoft...   Code   Custom code   En   Gptq   Phi-msft   Quantized   Region:us   Safetensors
Model Card on HF ๐Ÿค—: https://huggingface.co/TheBloke/phi-2-GPTQ 

Phi 2 GPTQ Benchmarks

nn.n% โ€” How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Sponsored by Nebius

Phi 2 GPTQ Parameters and Internals

Model Type 
text-generation, nlp, code
Use Cases 
Areas:
research purposes
Applications:
QA format, chat format, code format
Primary Use Cases:
research, exploration of safety challenges
Limitations:
potential for generating inaccurate code and facts, limited code scope to common packages, unreliable instructions adherence, language understanding mostly standard English, potential societal biases
Additional Notes 
Intended for research purposes. Not tested for production-level applications.
Supported Languages 
en (standard English)
Training Details 
Data Sources:
AOAI GPT-3.5, Falcon RefinedWeb, SlimPajama
Data Volume:
1.4T tokens
Methodology:
next-word prediction
Context Length:
2048
Training Time:
14 days
Hardware Used:
96xA100-80G
Model Architecture:
Transformer-based model
Input Output 
Accepted Modalities:
text
LLM NamePhi 2 GPTQ
Repository ๐Ÿค—https://huggingface.co/TheBloke/phi-2-GPTQ 
Model NamePhi 2
Model CreatorMicrosoft
Base Model(s)  Phi 2   microsoft/phi-2
Model Size600.4m
Required VRAM1.8โ€ฏGB
Updated2025-01-01
MaintainerTheBloke
Model Typephi-msft
Model Files  1.8 GB
Supported Languagesen
GPTQ QuantizationYes
Quantization Typegptq
Model ArchitecturePhiForCausalLM
Licenseother
Model Max Length2048
Transformers Version4.37.0.dev0
Tokenizer ClassCodeGenTokenizer
Vocabulary Size51200
Torch Data Typefloat16
Activation Functiongelu_new

Best Alternatives to Phi 2 GPTQ

Best Alternatives
Context / RAM
Downloads
Likes
Phi2 OSST GPTQ2K / 1.8โ€‰GB221
Phi 2 Orange GPTQ0K / 1.8โ€‰GB254
Dolphin 2 6 Phi 2 GPTQ0K / 1.8โ€‰GB282
LMCocktail Phi 2 V1 GPTQ0K / 1.8โ€‰GB231
Note: green Score (e.g. "73.2") means that the model is better than TheBloke/phi-2-GPTQ.

Rank the Phi 2 GPTQ 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  
โ—โ—โ—โ—
Factuality and Completeness of Knowledge  
โ—โ—โ—โ—
Censorship and Alignment  
โ—โ—โ—โ—
Data Analysis and Insight Generation  
โ—โ—โ—โ—
Text Generation  
โ—โ—โ—โ—
Text Summarization and Feature Extraction  
โ—โ—โ—โ—
Code Generation  
โ—โ—โ—โ—
Multi-Language Support and Translation  
โ—โ—โ—โ—

What open-source LLMs or SLMs are you in search of? 40649 in total.

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