What are the hardware requirements for ALMA 13B R?

ALMA 13B R requires approximately 26 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed ALMA 13B R and how large is it?

ALMA 13B R is developed by haoranxu, a model with 13b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does ALMA 13B R perform on standard benchmarks?

ALMA 13B R has the following published scores: MMLU 49.52. Compare against reference models on this page or on the LLM Explorer leaderboards.

ALMA 13B R by haoranxu — VRAM 26GB

Name: ALMA 13B R
Author: haoranxu

ALMA 13B R is an open-source language model by haoranxu. Features: 13b LLM, VRAM: 26GB, License: mit, LLM Explorer Score: 0.18, Arc: 55.6, HellaSwag: 79.5, MMLU: 49.5.

Arxiv:2309.11674 Arxiv:2401.08417 Base model:finetune:haoranxu/a... Base model:haoranxu/alma-13b-p... Deploy:azure Endpoints compatible Llama Lora Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/haoranxu/ALMA-13B-R

ALMA 13B R Parameters and Internals

Model Type

machine translation

Additional Notes

ALMA-7B-Pretrain and ALMA-13B-Pretrain are only monolingually fine-tuned and are not translation models independently.

Supported Languages

Chinese (High), English (High)

Training Details

Data Sources:

Human-Written Parallel Data, Triplet Preference Data

Methodology:

LoRA fine-tuning with Contrastive Preference Optimization

Input Output

Input Format:

Source sentence in prompt template

Accepted Modalities:

text

Output Format:

Translated output text

LLM Name	ALMA 13B R
Repository 🤗	https://huggingface.co/haoranxu/ALMA-13B-R
Base Model(s)	ALMA 13B Pretrain haoranxu/ALMA-13B-Pretrain
Model Size	13b
Required VRAM	26 GB
Updated	2026-07-11
Maintainer	haoranxu
Model Files	0.0 GB 5.0 GB: 1-of-6 5.0 GB: 2-of-6 5.0 GB: 3-of-6 4.9 GB: 4-of-6 4.9 GB: 5-of-6 1.2 GB: 6-of-6
Model Architecture	AutoModelForCausalLM
License	mit
Is Biased	none
Tokenizer Class	LlamaTokenizer
Padding Token	<unk>
PEFT Type	LORA
LoRA Model	Yes
PEFT Target Modules	down_proj
LoRA Alpha	32
LoRA Dropout	0.05
R Param	16

Rank the ALMA 13B R 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 ALMA 13B R

Best Alternatives	Context / RAM	Downloads	Likes
Small Tiny V1	0K / 0.1 GB	10	0
...ama2 13B No Robots Alpaca Lora	0K / 4 GB	58	11
Llava V1.5 13B S2 Lora	0K / 0.1 GB	9	2
Llama 2 13B Chat Hf GPT 4 80K	0K / 26 GB	63	0
Llava Chart 13b Lora	0K / 0.1 GB	4	2
ChartLlama 13B	0K / 0.1 GB	11	20
Orca 2 13B Pygmalion LoRA	0K / 0.5 GB	12	2
Leyley 13B Lora	0K / 4 GB	4	4
GTS Lewd 13B V0.2	0K / 52.3 GB	4	1
Testing Orca	0K / 0.5 GB	15	1