What are the hardware requirements for Galactica 30B Evol Instruct 70K GPTQ?

Galactica 30B Evol Instruct 70K GPTQ requires approximately 16.3 GB of VRAM and supports a context window of 2K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Galactica 30B Evol Instruct 70K GPTQ and how large is it?

Galactica 30B Evol Instruct 70K GPTQ is developed by TheBloke, a model with 30b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

Where can I download or evaluate Galactica 30B Evol Instruct 70K GPTQ?

Galactica 30B Evol Instruct 70K GPTQ is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Galactica 30B Evol Instruct 70K GPTQ by TheBloke — VRAM 16.3GB, 2K context

Name: Galactica 30B Evol Instruct 70K GPTQ
Author: TheBloke

Galactica 30B Evol Instruct 70K GPTQ is an open-source language model by TheBloke. Features: 30b LLM, VRAM: 16.3GB, Context: 2K, License: other, Quantized, Instruction-Based, LLM Explorer Score: 0.07.

Arxiv:2304.12244 4-bit Alpaca Dataset:wizardlm/wizardlm evol... Galactica Gptq Instruct Opt Quantized Region:us Safetensors Wizardlm

Model Card on HF 🤗: https://huggingface.co/TheBloke/galactica-30B-evol-instruct-70K-GPTQ

Galactica 30B Evol Instruct 70K GPTQ Parameters and Internals

Model Type

text generation

Use Cases

Areas:

Research

Applications:

Scientific computing

Primary Use Cases:

Research in language models applied to scientific domains

Limitations:

Not suitable for production use without safeguards due to potential inaccuracies

Considerations:

Ensure proper verification of generated information for scientific accuracy.

Additional Notes

Quantised to 4bit for improved efficiency.

Training Details

Data Sources:

WizardLM/WizardLM_evol_instruct_70k

Data Volume:

106 billion tokens

Methodology:

Fine-tuned using the Evol-Instruct framework.

Context Length:

1024

Training Time:

196 hours using 16 A100 80GB GPUs

Hardware Used:

16 A100 80GB GPUs

Model Architecture:

GPT-Q (quantized to 4bit)

Input Output

Input Format:

### Instruction: prompt ### Response:

Accepted Modalities:

text

Output Format:

Text

Performance Tips:

Use device='cuda:0' for better performance.

LLM Name	Galactica 30B Evol Instruct 70K GPTQ
Repository 🤗	https://huggingface.co/TheBloke/galactica-30B-evol-instruct-70K-GPTQ
Base Model(s)	...alactica 30B Evol Instruct 70K GeorgiaTechResearchInstitute/galactica-30b-evol-instruct-70k
Model Size	30b
Required VRAM	16.3 GB
Updated	2026-06-30
Maintainer	TheBloke
Model Type	opt
Instruction-Based	Yes
Model Files	16.3 GB
GPTQ Quantization	Yes
Quantization Type	gptq
Model Architecture	OPTForCausalLM
License	other
Context Length	2048
Model Max Length	2048
Transformers Version	4.29.2
Tokenizer Class	PreTrainedTokenizerFast
Vocabulary Size	50000
Torch Data Type	bfloat16
Activation Function	gelu

Rank the Galactica 30B Evol Instruct 70K GPTQ 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 Galactica 30B Evol Instruct 70K GPTQ

Best Alternatives	Context / RAM	Downloads	Likes
...alactica 30B Evol Instruct 70K	2K / 60.1 GB	57	23

Note: green Score (e.g. "73.2") means that the model is better than TheBloke/galactica-30B-evol-instruct-70K-GPTQ.