What are the hardware requirements for Gpt2 Large Conversational?

Gpt2 Large Conversational requires approximately 1.6 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Gpt2 Large Conversational and how large is it?

Gpt2 Large Conversational is developed by Locutusque, a model with 774m parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does Gpt2 Large Conversational perform on standard benchmarks?

Gpt2 Large Conversational has the following published scores: MMLU 26.33. Compare against reference models on this page or on the LLM Explorer leaderboards.

Gpt2 Large Conversational by Locutusque — VRAM 1.6GB

Name: Gpt2 Large Conversational
Author: Locutusque

Gpt2 Large Conversational is an open-source language model by Locutusque. Features: 774m LLM, VRAM: 1.6GB, License: openrail, Instruction-Based, LLM Explorer Score: 0.12, Arc: 27, HellaSwag: 45, MMLU: 26.3, GSM8K: 0.1.

Dataset:crumb/clean-instruct-4... Dataset:locutusque/columnedcha... Doi:10.57967/hf/1215 En Endpoints compatible Gpt2 Instruct Pytorch Region:us Safetensors

Model Card on HF 🤗: https://huggingface.co/Locutusque/gpt2-large-conversational

Gpt2 Large Conversational Parameters and Internals

Model Type

Language Modeling

Use Cases

Areas:

chatbots, virtual assistants, customer service applications

Applications:

text-based and voice-based interfaces

Primary Use Cases:

Generating Conversational Responses

Limitations:

Limited training time on a weak computer, Produces irrelevant or nonsensical responses, Not fine-tuned to remember chat history, Unable to provide follow-up responses, Does not know the answer to many questions

Additional Notes

Model is deprecated; a better performing version is available.

Supported Languages

en (proficient)

Training Details

Data Sources:

Locutusque/ColumnedChatCombined, crumb/Clean-Instruct-440k

Data Volume:

550,000 examples

Methodology:

Fine-tuned using maximum likelihood estimation

Training Time:

687,500 steps

Hardware Used:

1x Nvidia Titan V

Model Architecture:

Transformer-based, multi-layered decoder-only with self-attention mechanisms.

Input Output

Input Format:

<|USER|> {user prompt} <|ASSISTANT|>

Accepted Modalities:

text

Output Format:

<|USER|> {dataset prompt} <|ASSISTANT|> {dataset output}

LLM Name	Gpt2 Large Conversational
Repository 🤗	https://huggingface.co/Locutusque/gpt2-large-conversational
Model Size	774m
Required VRAM	1.6 GB
Updated	2026-08-04
Maintainer	Locutusque
Model Type	gpt2
Instruction-Based	Yes
Model Files	1.6 GB 1.6 GB
Supported Languages	en
Model Architecture	GPT2LMHeadModel
License	openrail
Model Max Length	1024
Transformers Version	4.30.1
Tokenizer Class	GPT2Tokenizer
Vocabulary Size	50260
Torch Data Type	bfloat16
Activation Function	gelu_new

Rank the Gpt2 Large Conversational 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 Gpt2 Large Conversational

Best Alternatives	Context / RAM	Downloads	Likes
Aira 2 774M	0K / 0 GB	171	3