Unakar1.5B Base by unakar

 ยป  All LLMs  ยป  unakar  ยป  Unakar1.5B Base   URL Share it on

  16k   Base Base model:finetune:unakar/una... Base model:unakar/unakar1.5b-b...   Custom code   En   Internlm2   Pytorch   Region:us   Scienceqa   Sharded   Textbooks
Model Card on HF ๐Ÿค—: https://huggingface.co/unakar/Unakar1.5B-base 

Unakar1.5B Base Benchmarks

nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Unakar1.5B Base (unakar/Unakar1.5B-base)

Unakar1.5B Base Parameters and Internals

Model Type 
text-generation
Additional Notes 
Base model without SFT and DPO alignment, not capable of instruction following.
Supported Languages 
en (high)
Training Details 
Data Sources:
huggingface open-source datasets, fineweb-edu, dclm, llama3-syne
Methodology:
Multi-stage pretraining with phases for general, math/code reasoning, and long context.
Training Time:
3 days
Hardware Used:
32 GPUs, data parallel, zero1
Model Architecture:
Based on LLaMA with modified mlp-ratio of 2.6875 for optimal TFLOPS.
LLM NameUnakar1.5B Base
Repository ๐Ÿค—https://huggingface.co/unakar/Unakar1.5B-base 
Base Model(s)  unakar/Unakar1.5B-base   unakar/Unakar1.5B-base
Model Size5b
Required VRAM3 GB
Updated2025-02-22
Maintainerunakar
Model Typeinternlm2
Model Files  2.0 GB: 1-of-2   1.0 GB: 2-of-2
Supported Languagesen
Model ArchitectureInternLM2ForCausalLM
Context Length16384
Model Max Length16384
Transformers Version4.25.1
Is Biased0
Tokenizer ClassInternLM2Tokenizer
Padding Token</s>
Vocabulary Size92544
Torch Data Typebfloat16

Rank the Unakar1.5B Base 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? 43470 in total.

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