What are the hardware requirements for Miqu Limarp 70B DPO Safefile?

Miqu Limarp 70B DPO Safefile requires approximately 38.4 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Miqu Limarp 70B DPO Safefile and how large is it?

Miqu Limarp 70B DPO Safefile is developed by NobodyExistsOnTheInternet, a model with 70b 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 Miqu Limarp 70B DPO Safefile?

Miqu Limarp 70B DPO Safefile is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Miqu Limarp 70B DPO Safefile by NobodyExistsOnTheInternet — VRAM 38.4GB

Name: Miqu Limarp 70B DPO Safefile
Author: NobodyExistsOnTheInternet

Miqu Limarp 70B DPO Safefile is an open-source language model by NobodyExistsOnTheInternet. Features: 70b LLM, VRAM: 38.4GB, LLM Explorer Score: 0.11.

4-bit Adapter Axolotl Base model:adapter:nobodyexist... Base model:nobodyexistsonthein... Bitsandbytes Dpo Finetuned Generated from trainer Llama Lora Peft Region:us Safetensors Sharded Tensorboard Tensorflow Trl

Model Card on HF 🤗: https://huggingface.co/NobodyExistsOnTheInternet/miqu-limarp-70b-dpo-safefile

Miqu Limarp 70B DPO Safefile Parameters and Internals

Model Type

AutoModelForCausalLM

Additional Notes

The model was trained using the axolotl library version 0.4.0 with PEFT 0.8.2.dev0, Transformers 4.37.0, Pytorch 2.0.1+cu117, Datasets 2.16.1, and Tokenizers 0.15.0.

Training Details

Data Sources:

NobodyExistsOnTheInternet/Fixed-FilteredTruthyDPO, NobodyExistsOnTheInternet/ToxicDPOqa, NobodyExistsOnTheInternet/Fixed-Distilabel-intel-orca-dpo-pairs, NobodyExistsOnTheInternet/Fixed-gutenberg-dpo-v0.1

Context Length:

768

Hardware Used:

multi-GPU, 8 devices

LLM Name	Miqu Limarp 70B DPO Safefile
Repository 🤗	https://huggingface.co/NobodyExistsOnTheInternet/miqu-limarp-70b-dpo-safefile
Base Model(s)	NobodyExistsOnTheInternet/3epoch-miqu-limarp NobodyExistsOnTheInternet/3epoch-miqu-limarp
Model Size	70b
Required VRAM	38.4 GB
Updated	2026-08-04
Maintainer	NobodyExistsOnTheInternet
Model Files	1.9 GB 5.0 GB: 1-of-8 5.0 GB: 2-of-8 5.0 GB: 3-of-8 5.0 GB: 4-of-8 5.0 GB: 5-of-8 5.0 GB: 6-of-8 5.0 GB: 7-of-8 3.4 GB: 8-of-8 0.0 GB
Model Architecture	Adapter
Is Biased	none
Tokenizer Class	LlamaTokenizer
Padding Token	<unk>
PEFT Type	LORA
LoRA Model	Yes
PEFT Target Modules	o_proj\|q_proj\|gate_proj\|down_proj\|k_proj\|v_proj\|up_proj
LoRA Alpha	64
LoRA Dropout	0.05
R Param	32

Rank the Miqu Limarp 70B DPO Safefile 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 Miqu Limarp 70B DPO Safefile

Best Alternatives	Context / RAM	Downloads	Likes
Llama 3.1 70B Abliterated Lora	0K / 1.7 GB	72	4
...aiga Llama3 70b Sft M1 D5 Lora	0K / 5.9 GB	0	1
Llama 3 70B Instruct Spider	0K / 141.9 GB	6	0
Airoboros 70B 3.3 Peft	0K / 0.4 GB	0	2
Llama3v1	0K / 0.1 GB	5	0
Xwin LM 70B V0.1 LORA	0K / 1.7 GB	6	1
Euryale 1.3 L2 70B LORA	0K / 1.7 GB	8	1
Miqu 1 70B Hermes2.5 Qlora	0K / 4.8 GB	7	4
Limarp Miqu 1 70B Qlora	0K / 1.7 GB	5	4
Limarp Miqu 1 70B Qlora	0K / 1.7 GB	2	4

Note: green Score (e.g. "73.2") means that the model is better than NobodyExistsOnTheInternet/miqu-limarp-70b-dpo-safefile.