Mallam 5B 4096 by mesolitica

 ยป  All LLMs  ยป  mesolitica  ยป  Mallam 5B 4096   URL Share it on

  Autotrain compatible   Endpoints compatible   Mistral   Ms   Region:us   Safetensors   Sharded   Tensorflow

Mallam 5B 4096 Benchmarks

nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Mallam 5B 4096 (mesolitica/mallam-5B-4096)

Mallam 5B 4096 Parameters and Internals

Model Type 
text generation
Additional Notes 
MaLLaM is specifically focused on Malaysian text, leveraging the Mistral architecture and fine-tuned for efficient text generation.
Supported Languages 
Malay (native)
Training Details 
Data Sources:
https://github.com/malaysia-ai/dedup-text-dataset/tree/main/pretrain-llm
Data Volume:
90B tokens
Context Length:
4096
Hardware Used:
5 nodes of 4x A100 80GB
Model Architecture:
Mistral
Input Output 
Input Format:
text
Accepted Modalities:
text
Output Format:
text
Performance Tips:
Use flash attention and 4-bit quantization for optimal performance.
LLM NameMallam 5B 4096
Repository ๐Ÿค—https://huggingface.co/mesolitica/mallam-5B-4096 
Model Size5b
Required VRAM10 GB
Updated2025-02-09
Maintainermesolitica
Model Typemistral
Model Files  4.9 GB: 1-of-3   4.8 GB: 2-of-3   0.3 GB: 3-of-3
Supported Languagesms
Model ArchitectureMistralForCausalLM
Context Length32768
Model Max Length32768
Transformers Version4.36.0.dev0
Tokenizer ClassPreTrainedTokenizerFast
Padding Token<pad>
Vocabulary Size32000
Torch Data Typebfloat16

Best Alternatives to Mallam 5B 4096

Best Alternatives
Context / RAM
Downloads
Likes
Mallam 5B 20K Instructions V232K / 10 GB201
Mallam 5B 20K Instructions32K / 10 GB701
Dummy Mistral 5B32K / 10 GB180
Note: green Score (e.g. "73.2") means that the model is better than mesolitica/mallam-5B-4096.

Rank the Mallam 5B 4096 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? 42848 in total.

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