What are the hardware requirements for Medllama3 V20?

Medllama3 V20 requires approximately 16.1 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Medllama3 V20 and how large is it?

Medllama3 V20 is developed by ProbeMedicalYonseiMAILab, a model with 8b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does Medllama3 V20 perform on standard benchmarks?

Medllama3 V20 has the following published scores: MMLU 67.2. Compare against reference models on this page or on the LLM Explorer leaderboards.

Medllama3 V20 by ProbeMedicalYonseiMAILab — VRAM 16.1GB

Name: Medllama3 V20
Author: ProbeMedicalYonseiMAILab

Medllama3 V20 is an open-source language model by ProbeMedicalYonseiMAILab. Features: 8b LLM, VRAM: 16.1GB, License: llama3, LLM Explorer Score: 0.2, Arc: 48.4, HellaSwag: 72.4, MMLU: 67.2, GSM8K: 46.6.

Autotrain compatible Conversational Endpoints compatible Llama Lora Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/ProbeMedicalYonseiMAILab/medllama3-v20

Medllama3 V20 Parameters and Internals

Model Type

LLM

Additional Notes

The model utilizes the following training hyperparameters: Lora Targets: "o_proj", "down_proj", "v_proj", "gate_proj", "up_proj", "k_proj", "q_proj".

Supported Languages

English ()

Training Details

Data Sources:

publically available medical data

Methodology:

fine-tuning

LLM Name	Medllama3 V20
Repository 🤗	https://huggingface.co/ProbeMedicalYonseiMAILab/medllama3-v20
Model Size	8b
Required VRAM	16.1 GB
Updated	2025-03-05
Maintainer	ProbeMedicalYonseiMAILab
Model Files	0.1 GB 4.0 GB: 1-of-5 3.9 GB: 2-of-5 3.9 GB: 3-of-5 3.2 GB: 4-of-5 1.1 GB: 5-of-5
Model Architecture	AutoModelForCausalLM
License	llama3
Is Biased	none
Tokenizer Class	PreTrainedTokenizerFast
Padding Token	<\|eot_id\|>
PEFT Type	LORA
LoRA Model	Yes
PEFT Target Modules	v_proj\|up_proj\|down_proj\|gate_proj\|k_proj\|o_proj\|q_proj
LoRA Alpha	32
LoRA Dropout	0.1
R Param	16

Rank the Medllama3 V20 capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

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

Best Alternatives to Medllama3 V20

Best Alternatives	Context / RAM	Downloads	Likes
Llama 3.1 8B Instuct Uz	128K / 16.1 GB	1000	15
Trillama 8B	8K / 16.1 GB	14	3
AutogenJune2000	8K / 33.8 GB	6	0
Llama3 8B	8K / 16.1 GB	7	0
NewMes V15	0K / 16.1 GB	409	18
Cve Model	0K / 0.3 GB	16	0
Shaista Pro	0K / 0 GB	10	0
...Datasets Layer 1 Decoder Fixed	0K / 0.7 GB	672	0
ConfTuner Ministral	0K / 16.1 GB	13	4
...llem V1 Smilebase Llama 3.1 8B	0K / 1.1 GB	13	0

Note: green Score (e.g. "73.2") means that the model is better than ProbeMedicalYonseiMAILab/medllama3-v20.