LLM EXPLORER
 LLM Explorer: A Curated Large Language Model Directory and Analytics
Dark Theme
  LLM  List    LLM Hosting    LLM Leaderboards    Blog    Newsfeed    Advertise

MediaTek Research Breeze 7B Instruct V1.0 Bnb 4bit Smashed by PrunaAI

 »  All LLMs  »  PrunaAI  »  MediaTek Research Breeze 7B Instruct V1.0 Bnb 4bit Smashed   URL Share it on

  15   
  4-bit   4bit   Autotrain compatible Base model:mediatek-research/b... Base model:quantized:mediatek-...   Bitsandbytes   Conversational   Endpoints compatible   Instruct   Mistral   Pruna-ai   Quantized   Region:us   Safetensors
Model Card on HF 🤗: https://huggingface.co/PrunaAI/MediaTek-Research-Breeze-7B-Instruct-v1_0-bnb-4bit-smashed 

MediaTek Research Breeze 7B Instruct V1.0 Bnb 4bit Smashed Benchmarks

LLME Score: 0.18269
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
MediaTek Research Breeze 7B Instruct V1.0 Bnb 4bit Smashed (PrunaAI/MediaTek-Research-Breeze-7B-Instruct-v1_0-bnb-4bit-smashed)

MediaTek Research Breeze 7B Instruct V1.0 Bnb 4bit Smashed Parameters and Internals

Additional Notes 
Calibration data used is WikiText. Metrics results vary with 'first', 'Sync', 'Async' mention for different run conditions.
Training Details 
Data Sources:
WikiText
Methodology:
Model compressed with llm-int8
Input Output 
Output Format:
safetensors
Performance Tips:
Ensure you use conditions similar to your final use case to assess performance benefits.
Release Notes 
Notes:
Model smashed and compressed with reduction in inference speed, memory, and energy consumption.
LLM NameMediaTek Research Breeze 7B Instruct V1.0 Bnb 4bit Smashed
Repository 🤗https://huggingface.co/PrunaAI/MediaTek-Research-Breeze-7B-Instruct-v1_0-bnb-4bit-smashed 
Base Model(s)  Breeze 7B Instruct V1.0   MediaTek-Research/Breeze-7B-Instruct-v1_0
Model Size7b
Required VRAM4.9 GB
Updated2025-01-20
MaintainerPrunaAI
Model Typemistral
Instruction-BasedYes
Model Files  4.9 GB
Quantization Type4bit
Model ArchitectureMistralForCausalLM
Context Length32768
Model Max Length32768
Transformers Version4.40.0
Tokenizer ClassLlamaTokenizer
Padding Token</s>
Vocabulary Size61952
Torch Data Typefloat16

Best Alternatives to MediaTek Research Breeze 7B Instruct V1.0 Bnb 4bit Smashed

Best Alternatives
Context / RAM
Downloads
Likes
...al Nemo Instruct 2407 Bnb 4bit1000K / 8.3 GB1286425
...tral 7B Instruct V0.3 Bnb 4bit32K / 4.1 GB15148517
Mistral 7B Instruct V0.2 Fp1632K / 14.4 GB320
Mistral 7B Instruct V0.2 4bit32K / 4.3 GB2291
...tral 7B Instruct V0.2 Bnb 4bit32K / 4.1 GB1141931
NaturalLM 7B Instruct32K / 14.5 GB1910
...eMistral 7B Instruct Ead 16bit32K / 14.5 GB380
Mistral 7B Instruct V0.2 8bit32K / 7.6 GB181
Mistral Turkish V232K / 14.4 GB30396
Minirewrite32K / 15.8 GB230
Note: green Score (e.g. "73.2") means that the model is better than PrunaAI/MediaTek-Research-Breeze-7B-Instruct-v1_0-bnb-4bit-smashed.

Rank the MediaTek Research Breeze 7B Instruct V1.0 Bnb 4bit Smashed Capabilities

🆘 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! 🌟

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? 41636 in total.

Email us: info@extractum.io. Our Privacy Policy  | Terms and Conditions  | Suggest an improvement.
Our Social Media →  
Original data from HuggingFace, OpenCompass and various public git repos.
Release v20241227
  Support LLM Explorer