Phi 3.5 Mini Instruct Onnx Web by onnx-community

 ยป  All LLMs  ยป  onnx-community  ยป  Phi 3.5 Mini Instruct Onnx Web   URL Share it on

  Arxiv:2404.14219   Arxiv:2407.13833   Code   Conversational   Instruct   Multilingual   Onnx   Phi3   Region:us   Transformers.js

Phi 3.5 Mini Instruct Onnx Web Benchmarks

nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Phi 3.5 Mini Instruct Onnx Web (onnx-community/Phi-3.5-mini-instruct-onnx-web)

Phi 3.5 Mini Instruct Onnx Web Parameters and Internals

Model Type 
text-generation
Use Cases 
Areas:
commercial use, research
Applications:
general purpose AI systems
Primary Use Cases:
memory/compute constrained environments, latency bound scenarios, strong reasoning, accelerate research on language and multimodal models
Limitations:
not designed for all downstream purposes
Considerations:
developed with mapping, measuring, and mitigating risks in mind
Supported Languages 
Arabic (supported), Chinese (supported), Czech (supported), Danish (supported), Dutch (supported), English (supported), Finnish (supported), French (supported), German (supported), Hebrew (supported), Hungarian (supported), Italian (supported), Japanese (supported), Korean (supported), Norwegian (supported), Polish (supported), Portuguese (supported), Russian (supported), Spanish (supported), Swedish (supported), Thai (supported), Turkish (supported), Ukrainian (supported)
Training Details 
Data Sources:
publicly available documents, synthetic data
Data Volume:
3.4T tokens
Methodology:
supervised fine-tuning, proximal policy optimization, and direct preference optimization
Context Length:
128000
Training Time:
10 days
Hardware Used:
512 H100-80G GPUs
Model Architecture:
dense decoder-only Transformer
Safety Evaluation 
Methodologies:
red-teaming, adversarial tests
Findings:
models may refuse undesirable outputs in English even in other languages, susceptible to longer multi-turn jailbreak techniques
Risk Categories:
misinformation, bias
Ethical Considerations:
invest in quality safety datasets across multiple languages
Responsible Ai Considerations 
Fairness:
language models can be unfair or offensive
Transparency:
developers should inform users they are interacting with an AI
Accountability:
developers should fine-tune models for specific uses
Mitigation Strategies:
apply responsible AI best practices
Input Output 
Input Format:
chat format
Accepted Modalities:
Text
Output Format:
Generated text
Performance Tips:
testing in specific AI applications is encouraged
Release Notes 
Version:
Phi-3.5-mini
Date:
August 2024
Notes:
update over June 2024 release, improved multilingual and reasoning capability
LLM NamePhi 3.5 Mini Instruct Onnx Web
Repository ๐Ÿค—https://huggingface.co/onnx-community/Phi-3.5-mini-instruct-onnx-web 
Updated2024-12-22
Maintaineronnx-community
Model Typephi3
Instruction-BasedYes
Model ArchitecturePhi3ForCausalLM
Licensemit
Context Length131072
Model Max Length131072
Transformers Version4.43.3
Tokenizer ClassLlamaTokenizer
Padding Token<|endoftext|>
Vocabulary Size32064
Torch Data Typebfloat16

Best Alternatives to Phi 3.5 Mini Instruct Onnx Web

Best Alternatives
Context / RAM
Downloads
Likes
Phi 3.5 Mini Instruct Onnx128K /  GB31722
Phi 3 Mini 128K Instruct Onnx128K /  GB411185
...Medium 128K Instruct Onnx Cuda128K /  GB46623
... Medium 128K Instruct Onnx Cpu128K /  GB48811
...i 3 Mini 128K Instruct Onnx Tf128K /  GB3991
...i 3 Mini 128K Instruct Ov Int4128K / 0 GB80
...um 128K Instruct Onnx Directml128K /  GB425
Model1128K / 0.8 GB80
...3 Mini 128K Instruct Asym Int4128K / 2.5 GB70
...i 3 Mini 128K Instruct Ov Int8128K / 0 GB231
Note: green Score (e.g. "73.2") means that the model is better than onnx-community/Phi-3.5-mini-instruct-onnx-web.

Rank the Phi 3.5 Mini Instruct Onnx Web 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? 40066 in total.

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