LLM EXPLORER 60,828 MODELS INDEXED

Phi 3 Mini 128K Instruct Ov Int4 by fakezeta

By fakezeta · 11 downloads
HF Repo ↗ Copy URL

Phi 3 Mini 128K Instruct Ov Int4 is an open-source language model by fakezeta. Features: LLM, VRAM: 2.5GB, Context: 128K, License: mit, Instruction-Based, LLM Explorer Score: 0.12.

  Conversational   Custom code   Endpoints compatible   Instruct   Openvino   Phi3   Region:us
Model Card on HF 🤗: https://huggingface.co/fakezeta/Phi-3-mini-128k-instruct-ov-int4 

Phi 3 Mini 128K Instruct Ov Int4 Parameters and Internals

Model Type 
dense, decoder-only, Transformer
Use Cases 
Areas:
commercial, research
Applications:
Memory/compute constrained environments, Latency bound scenarios, Strong reasoning (code, math, and logic)
Primary Use Cases:
commercial applications, research use in English
Limitations:
Not evaluated for all downstream purposes, Performance primarily in English
Considerations:
Developers should evaluate for accuracy, safety, and fairness, especially in high-risk scenarios.
Additional Notes 
Integration with transformers development version 4.40.0, using flash attention by default. Optimized for various iterations of GPU and CPU hardware.
Supported Languages 
English (primary)
Training Details 
Data Sources:
Publicly available documents, synthetic data, high-quality educational data, code
Data Volume:
3.3 trillion tokens
Methodology:
Supervised fine-tuning and Direct Preference Optimization
Context Length:
4000
Training Time:
7 days
Hardware Used:
512 H100-80G GPUs
Model Architecture:
dense decoder-only Transformer
Safety Evaluation 
Risk Categories:
misinformation, bias, offensiveness
Ethical Considerations:
Developers should ensure the model complies with relevant laws and regulations.
Responsible Ai Considerations 
Fairness:
Evaluated for instructional following and safety measures.
Transparency:
Developers should inform users they are interacting with an AI system.
Accountability:
Developers are responsible for their specific use cases complying with laws.
Mitigation Strategies:
Use available safety classifiers or custom solutions.
Input Output 
Input Format:
Chat format with <|user|> and <|assistant|> tags
Accepted Modalities:
text
Output Format:
Generated text in response to input prompts
Performance Tips:
For NVIDIA V100 or earlier, use attn_implementation="eager"
LLM NamePhi 3 Mini 128K Instruct Ov Int4
Repository 🤗https://huggingface.co/fakezeta/Phi-3-mini-128k-instruct-ov-int4 
Required VRAM2.5 GB
Updated2026-08-04
Maintainerfakezeta
Model Typephi3
Instruction-BasedYes
Model Files  2.5 GB
Model ArchitecturePhi3ForCausalLM
Licensemit
Context Length131072
Model Max Length131072
Transformers Version4.39.3
Tokenizer ClassLlamaTokenizer
Padding Token<|endoftext|>
Vocabulary Size32064
Torch Data Typebfloat16

Quantized Models of the Phi 3 Mini 128K Instruct Ov Int4

Model
Likes
Downloads
VRAM
...28K Instruct Ov Fp16 Int4 Asym092 GB

Best Alternatives to Phi 3 Mini 128K Instruct Ov Int4

Best Alternatives
Context / RAM
Downloads
Likes
Phi 4 Mini Instruct ONNX GQA128K /  GB476
Phi 3.5 Mini Instruct Onnx128K /  GB13839
Phi 3.5 Mini Instruct Onnx Web128K /  GB40715
Phi 3 Mini 128K Instruct Onnx128K /  GB86193
...i 3 Mini 128K Instruct Ov Int4128K / 2 GB80
...Medium 128K Instruct Onnx Cuda128K /  GB8224
... Medium 128K Instruct Onnx Cpu128K /  GB8414
...um 128K Instruct Onnx Directml128K /  GB896
...3 Mini 128K Instruct Asym Int4128K / 2.5 GB1220
Meeting Model Phi3 128k128K / 7.7 GB50
Note: green Score (e.g. "73.2") means that the model is better than fakezeta/Phi-3-mini-128k-instruct-ov-int4.