LLM EXPLORER 60,783 MODELS INDEXED

Llama 3.1 Nemotron 70B Instruct HF 8.0bpw 8hb EXL2 by bigstorm

By bigstorm · 12 downloads
HF Repo ↗ Copy URL

Llama 3.1 Nemotron 70B Instruct HF 8.0bpw 8hb EXL2 is an open-source language model by bigstorm. Features: 70b LLM, VRAM: 64.3GB, Context: 128K, License: llama3.1, Quantized, Instruction-Based, LLM Explorer Score: 0.15.

  Arxiv:2310.05344   Arxiv:2311.09528   Arxiv:2406.08673   Arxiv:2410.01257   8-bit Base model:meta-llama/llama-3.... Base model:quantized:meta-llam...   Conversational   Dataset:nvidia/helpsteer2   En   Exl2   Instruct   Llama   Llama3.1   Nvidia   Quantized   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/bigstorm/Llama-3.1-Nemotron-70B-Instruct-HF-8.0bpw-8hb-exl2 

Llama 3.1 Nemotron 70B Instruct HF 8.0bpw 8hb EXL2 Parameters and Internals

Model Type 
text generation
Use Cases 
Areas:
General-domain instruction following
Limitations:
Not tuned for performance in specialized domains like math
Additional Notes 
This model is a demonstration of NVIDIA's techniques for improving helpfulness in general-domain instruction following. Evaluation results might differ from when the model is evaluated in NeMo-Aligner.
Supported Languages 
en (Native)
Training Details 
Data Sources:
nvidia/HelpSteer2
Methodology:
Trained using RLHF (specifically, REINFORCE), HelpSteer2-Preference prompts and Llama-3.1-Nemotron-70B-Instruct as initial policy.
Context Length:
128000
Model Architecture:
Transformer
Input Output 
Input Format:
String
Accepted Modalities:
Text
Output Format:
String
LLM NameLlama 3.1 Nemotron 70B Instruct HF 8.0bpw 8hb EXL2
Repository 🤗https://huggingface.co/bigstorm/Llama-3.1-Nemotron-70B-Instruct-HF-8.0bpw-8hb-exl2 
Base Model(s)  meta-llama/Llama-3.1-70B-Instruct   meta-llama/Llama-3.1-70B-Instruct
Model Size70b
Required VRAM64.3 GB
Updated2026-07-28
Maintainerbigstorm
Model Typellama
Instruction-BasedYes
Model Files  8.5 GB: 1-of-8   8.4 GB: 2-of-8   8.4 GB: 3-of-8   8.5 GB: 4-of-8   8.6 GB: 5-of-8   8.4 GB: 6-of-8   8.4 GB: 7-of-8   5.1 GB: 8-of-8
Supported Languagesen
Quantization Typeexl2
Model ArchitectureLlamaForCausalLM
Licensellama3.1
Context Length131072
Model Max Length131072
Transformers Version4.40.0
Tokenizer ClassPreTrainedTokenizerFast
Vocabulary Size128256
Torch Data Typebfloat16

Best Alternatives to Llama 3.1 Nemotron 70B Instruct HF 8.0bpw 8hb EXL2

Best Alternatives
Context / RAM
Downloads
Likes
...B Instruct Gradient 1048K 8bit1024K / 75 GB415
...B Instruct Gradient 1048K 4bit1024K / 39.7 GB253
...B Instruct Gradient 1048K 2bit1024K / 21.9 GB252
...0B Instruct Gradient 262K 4bit256K / 39.7 GB253
...0B Instruct Gradient 262K 8bit256K / 75 GB262
...0B Instruct Gradient 262K 2bit256K / 21.9 GB201
... Gradient 262K 2.25bpw H6 EXL2256K / 22.2 GB80
...t Gradient 262K 4.0bpw H6 EXL2256K / 37.2 GB71
...t Gradient 262K 2.4bpw H6 EXL2256K / 23.5 GB60
...t Gradient 262K 3.5bpw H6 EXL2256K / 32.9 GB60