What are the hardware requirements for Llama 3.1 Cantonese 8B Instruct?

Llama 3.1 Cantonese 8B Instruct requires approximately 16.1 GB of VRAM and supports a context window of 128K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Llama 3.1 Cantonese 8B Instruct and how large is it?

Llama 3.1 Cantonese 8B Instruct is developed by lordjia, 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 Llama 3.1 Cantonese 8B Instruct perform on standard benchmarks?

Llama 3.1 Cantonese 8B Instruct has the following published scores: MMLU 63.1. Compare against reference models on this page or on the LLM Explorer leaderboards.

Llama 3.1 Cantonese 8B Instruct by lordjia — VRAM 16.1GB, 128K context

Name: Llama 3.1 Cantonese 8B Instruct
Author: lordjia

Llama 3.1 Cantonese 8B Instruct is an open-source language model by lordjia. Features: 8b LLM, VRAM: 16.1GB, Context: 128K, License: llama3.1, Quantized, Instruction-Based, LLM Explorer Score: 0.16, Arc: 61.3, HellaSwag: 80.2, MMLU: 63.1, GSM8K: 44.4.

Cantonese Chat Conversational Dataset:jed351/cantonese-wikip... Dataset:lordjia/cantonese engl... En Endpoints compatible Gguf Instruct Llama Llama3.1 Q4 Quantized Region:us Safetensors Sharded Tensorflow Zh

Model Card on HF 🤗: https://huggingface.co/lordjia/Llama-3.1-Cantonese-8B-Instruct

Llama 3.1 Cantonese 8B Instruct Parameters and Internals

LLM Name	Llama 3.1 Cantonese 8B Instruct
Repository 🤗	https://huggingface.co/lordjia/Llama-3.1-Cantonese-8B-Instruct
Model Size	8b
Required VRAM	16.1 GB
Updated	2026-08-08
Maintainer	lordjia
Model Type	llama
Instruction-Based	Yes
Model Files	4.7 GB 5.0 GB: 1-of-4 5.0 GB: 2-of-4 4.9 GB: 3-of-4 1.2 GB: 4-of-4
Supported Languages	en zh
GGUF Quantization	Yes
Quantization Type	q4\|gguf
Model Architecture	LlamaForCausalLM
License	llama3.1
Context Length	131072
Model Max Length	131072
Transformers Version	4.43.3
Tokenizer Class	PreTrainedTokenizerFast
Padding Token	<\|eot_id\|>
Vocabulary Size	128256
Torch Data Type	bfloat16

Rank the Llama 3.1 Cantonese 8B Instruct 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 Llama 3.1 Cantonese 8B Instruct

Best Alternatives	Context / RAM	Downloads	Likes
...truct Gradient 1048K IMat GGUF	1024K / 2 GB	650	6
...B Instruct Gradient 1048K GGUF	1024K / 3.2 GB	445	3
Llama 3 8B Instruct 262K GGUF	256K / 3.2 GB	208	2
Grok Oss Apollyon 8B	128K / 16.1 GB	3449	8
Grok Oss Revenant 8B	128K / 16.1 GB	324	0
Llama 3.1 8B Instruct GGUF	128K / 2.2 GB	17956	41
... 8B Instruct Reasoner 1o1 V0.3	128K / 16.1 GB	322	8
...tion Sec 8B Instruct Q8 0 GGUF	128K / 8.5 GB	905	6
Llama 3.1 TAIDE R1 8B Chat	128K / 17.1 GB	1383	23
Nsfw Plz Gguf Me	128K / 16.1 GB	3	3

Note: green Score (e.g. "73.2") means that the model is better than lordjia/Llama-3.1-Cantonese-8B-Instruct.