LLM EXPLORER 59,420 MODELS INDEXED

MMedLM2 by Henrychur

By Henrychur · 12 downloads
HF Repo ↗ Copy URL

MMedLM2 is an open-source language model by Henrychur. Features: LLM, VRAM: 30.9GB, Context: 32K, License: apache-2.0, LLM Explorer Score: 0.11.

  Arxiv:2402.13963   Custom code   Dataset:henrychur/mmedc   En   Es   Feature-extraction   Fr   Internlm   Ja   Medical   Pytorch   Region:us   Ru   Sharded   Zh
Model Card on HF 🤗: https://huggingface.co/Henrychur/MMedLM2 

MMedLM2 Parameters and Internals

Model Type 
foundation model, multilingual medical model
Additional Notes 
The model is further pre-trained on a comprehensive multilingual medical corpus (MMedC).
Supported Languages 
en (high), zh (high), ja (high), fr (high), ru (high), es (high)
Training Details 
Data Sources:
MMedC
Data Volume:
25.5 billion tokens
Context Length:
2048
Input Output 
Input Format:
multilingual text
Accepted Modalities:
text
Output Format:
multilingual text
Performance Tips:
Loading with torch_dtype=torch.float16 improves performance.
Release Notes 
Date:
2024.2.21
Notes:
Pre-print paper released, Dive into our findings.
Date:
2024.2.20
Notes:
Released MMedLM and MMedLM 2
LLM NameMMedLM2
Repository 🤗https://huggingface.co/Henrychur/MMedLM2 
Required VRAM30.9 GB
Updated2026-07-27
MaintainerHenrychur
Model Typeinternlm
Model Files  9.8 GB: 1-of-4   9.8 GB: 2-of-4   9.8 GB: 3-of-4   1.5 GB: 4-of-4
Supported Languagesen zh ja fr ru es
Model ArchitectureInternLM2ForCausalLM
Licenseapache-2.0
Context Length32768
Model Max Length32768
Transformers Version4.28.1
Is Biased0
Tokenizer ClassInternLMTokenizer
Padding Token</s>
Vocabulary Size92544
Torch Data Typefloat32

Best Alternatives to MMedLM2

Best Alternatives
Context / RAM
Downloads
Likes
ShareCaptioner Video32K / 17.2 GB9519
Songcomposer Sft32K / 16.7 GB6816
Songcomposer Pretrain32K / 16.7 GB265
Internlm2 5 Step Prover8K / 15.4 GB4885
Internlm2 Step Prover8K / 15.4 GB3623
Note: green Score (e.g. "73.2") means that the model is better than Henrychur/MMedLM2.