What are the hardware requirements for ALMA 13B Pretrain?

ALMA 13B Pretrain requires approximately 52.1 GB of VRAM and supports a context window of 4K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

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

ALMA 13B Pretrain 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 Pretrain perform on standard benchmarks?

ALMA 13B Pretrain has the following published scores: MMLU 50.31. Compare against reference models on this page or on the LLM Explorer leaderboards.

ALMA 13B Pretrain by haoranxu — VRAM 52.1GB, 4K context

Name: ALMA 13B Pretrain
Author: haoranxu

ALMA 13B Pretrain is an open-source language model by haoranxu. Features: 13b LLM, VRAM: 52.1GB, Context: 4K, License: mit, LLM Explorer Score: 0.15, Arc: 56.9, HellaSwag: 80.2, MMLU: 50.3, GSM8K: 8.9.

Arxiv:2309.11674 Arxiv:2401.08417 Base model:finetune:meta-llama... Base model:meta-llama/llama-2-... Deploy:azure Endpoints compatible Llama Pytorch Region:us Sharded

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

ALMA 13B Pretrain Benchmarks

ARC: 56.91 vs 96.7 (so35)^-41.1%

HellaSwag: 80.15 vs 95.3 (gpt4)^-15.9%

MMLU: 50.31 vs 88.3 (so35)^-43%

GSM8K: 8.87 vs 96.4 (so35)^-90.8%

LLME Score: 0.14961

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

What is the LLM Explorer Rank (Score)

ALMA 13B Pretrain (haoranxu/ALMA-13B-Pretrain)

🌟 Advertise your project 🚀

ALMA 13B Pretrain Parameters and Internals

Model Type

Translation

Use Cases

Areas:

Translation, Research

Applications:

Machine Translation

Primary Use Cases:

Translating Chinese to English

Considerations:

For usage with LoRA models.

Additional Notes

ALMA-R incorporates Contrastive Preference Optimization (CPO) for improved performance.

Supported Languages

Monolingual Data (Multi-lingual), Parallel Data (High-quality parallel data)

Training Details

Data Sources:

20B monolingual tokens, high-quality parallel data, triplet preference data

Data Volume:

20B tokens for 7B model and 12B tokens for 13B model

Methodology:

Two-step fine-tuning with monolingual data followed by parallel data; Further optimized with Contrastive Preference Optimization (CPO)

Context Length:

Model Architecture:

LLaMA based, fine-tuned with LoRA

Input Output

Input Format:

Text input

Accepted Modalities:

text

Output Format:

Text output

Performance Tips:

Use LoRA models together with Base Models for intended performance.

Release Notes

Version:

ALMA-13B-LoRA

Date:

unknown

Notes:

Full-weight Fine-tune LLaMA-2-7B on 12B monolingual tokens and then LoRA fine-tune on human-written parallel data

Version:

ALMA-13B-R

Date:

unknown

Notes:

Further LoRA fine-tuning upon ALMA-13B-LoRA with contrastive preference optimization

LLM Name	ALMA 13B Pretrain
Repository 🤗	https://huggingface.co/haoranxu/ALMA-13B-Pretrain
Base Model(s)	Llama 2 13B Hf meta-llama/Llama-2-13b-hf
Model Size	13b
Required VRAM	52.1 GB
Updated	2026-05-08
Maintainer	haoranxu
Model Type	llama
Model Files	10.0 GB: 1-of-6 9.9 GB: 2-of-6 9.9 GB: 3-of-6 9.9 GB: 4-of-6 9.9 GB: 5-of-6 2.5 GB: 6-of-6
Model Architecture	LlamaForCausalLM
License	mit
Context Length	4096
Model Max Length	4096
Transformers Version	4.30.0.dev0
Tokenizer Class	LlamaTokenizer
Beginning of Sentence Token	<s>
End of Sentence Token	</s>
Unk Token	<unk>
Vocabulary Size	32000
Torch Data Type	float32

Quantized Models of the ALMA 13B Pretrain

Model	Likes	Downloads	VRAM
ALMA 13B Pretrain GGUF	12	333	5 GB
ALMA 13B Pretrain GPTQ	1	11	7 GB
ALMA 13B Pretrain AWQ	1	9	7 GB

Best Alternatives to ALMA 13B Pretrain

Best Alternatives	Context / RAM	Downloads	Likes
Luminaura RP 13B	128K / 26 GB	15	1
Yarn Llama 2 13B 128K	128K / 26 GB	1657	113
Agent Llama2 13B 80K	80K / 26.4 GB	5	0
Chat Llama2 13B 80K	80K / 52.8 GB	11	0
LongAlign 13B 64K	64K / 26 GB	53	13
LongAlign 13B 64K	64K / 26 GB	11	13
LongAlign 13B 64K Base	64K / 26 GB	41	3
LongAlign 13B 64K Base	64K / 26 GB	6	3
Openbuddy Llama2 13B V15p1 64K	64K / 26.1 GB	7	4
Yarn Llama 2 13B 64K	64K / 26 GB	1865	18

Rank the ALMA 13B Pretrain Capabilities

🆘 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! 🌟

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? 53999 in total.

Email us: info@extractum.io. Our Privacy Policy | Terms and Conditions | Suggest an improvement.

Our Social Media →

Original data from HuggingFace, OpenCompass and various public git repos.

Check out Ag3ntum — our secure, self-hosted AI agent for server management.

Release v20260328a

Support LLM Explorer

ALMA 13B Pretrain by haoranxu