Medicine LLM by AdaptLLM

 ยป  All LLMs  ยป  AdaptLLM  ยป  Medicine LLM   URL Share it on

  Arxiv:2309.09530   Arxiv:2406.14491   Arxiv:2411.19930   Autotrain compatible   Biology   Dataset:eleutherai/pile   Dataset:gair/lima   Dataset:open-orca/openorca Dataset:wizardlm/wizardlm evol...   En   Endpoints compatible   Instruct   Llama   Medical   Pytorch   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF ๐Ÿค—: https://huggingface.co/AdaptLLM/medicine-LLM 

Medicine LLM Benchmarks

nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Medicine LLM (AdaptLLM/medicine-LLM)

Medicine LLM Parameters and Internals

Model Type 
text generation
Use Cases 
Areas:
biomedicine, finance, law
Applications:
question answering, reading comprehension, domain-specific applications
Primary Use Cases:
Domain-specific task handling in biomedicine, finance, and law
Limitations:
Reduced prompting ability if not using the reading comprehension approach
Considerations:
Pre-filled instructions are tailored for non-aligned models; different formats are required for chat capabilities
Additional Notes 
The 13B version scales up LLaMA-1-13B for larger-scale models with positive results
Training Details 
Data Sources:
Open-Orca/OpenOrca, GAIR/lima, WizardLM/WizardLM_evol_instruct_V2_196k, EleutherAI/pile
Methodology:
Continual pre-training on domain-specific corpora, using reading comprehension texts to improve prompting performance
Context Length:
2048
Model Architecture:
Based on LLaMA-1-7B, adapted for domain-specific tasks
Input Output 
Input Format:
text input in question form
Accepted Modalities:
text
Output Format:
text answer or completion
Performance Tips:
Use the reading comprehension transformation to boost prompting performance in domain-specific tasks
Release Notes 
Version:
2nd version
Date:
2024/6/21
Notes:
Released at Instruction-Pretrain, effective for both pre-training and continual pre-training
LLM NameMedicine LLM
Repository ๐Ÿค—https://huggingface.co/AdaptLLM/medicine-LLM 
Model Size6.7b
Required VRAM8.8 GB
Updated2024-12-23
MaintainerAdaptLLM
Model Typellama
Instruction-BasedYes
Model Files  0.8 GB: 1-of-33   0.8 GB: 2-of-33   0.8 GB: 3-of-33   0.8 GB: 4-of-33   0.8 GB: 5-of-33   0.8 GB: 6-of-33   0.8 GB: 7-of-33   0.8 GB: 8-of-33   0.8 GB: 9-of-33   0.8 GB: 10-of-33   0.8 GB: 11-of-33   0.8 GB: 12-of-33   0.8 GB: 13-of-33   0.8 GB: 14-of-33   0.8 GB: 15-of-33   0.8 GB: 16-of-33   0.8 GB: 17-of-33   0.8 GB: 18-of-33   0.8 GB: 19-of-33   0.8 GB: 20-of-33   0.8 GB: 21-of-33   0.8 GB: 22-of-33   0.8 GB: 23-of-33   0.8 GB: 24-of-33   0.8 GB: 25-of-33   0.8 GB: 26-of-33   0.8 GB: 27-of-33   0.8 GB: 28-of-33   0.8 GB: 29-of-33   0.8 GB: 30-of-33   0.8 GB: 31-of-33   0.8 GB: 32-of-33   0.5 GB: 33-of-33   0.8 GB: 1-of-33   0.8 GB: 2-of-33   0.8 GB: 3-of-33   0.8 GB: 4-of-33   0.8 GB: 5-of-33   0.8 GB: 6-of-33   0.8 GB: 7-of-33   0.8 GB: 8-of-33   0.8 GB: 9-of-33   0.8 GB: 10-of-33   0.8 GB: 11-of-33
Supported Languagesen
Model ArchitectureLLaMAForCausalLM
Transformers Version4.27.0.dev0
Vocabulary Size32001
Torch Data Typefloat16

Quantized Models of the Medicine LLM

Model
Likes
Downloads
VRAM
Medicine LLM GGUF156752 GB
Medicine LLM GPTQ12293 GB
Medicine LLM AWQ3383 GB

Best Alternatives to Medicine LLM

Best Alternatives
Context / RAM
Downloads
Likes
Finance LLM0K / 26.1 GB298110
Law LLM0K / 26.1 GB16258
Note: green Score (e.g. "73.2") means that the model is better than AdaptLLM/medicine-LLM.

Rank the Medicine LLM 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? 40123 in total.

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