Quantization made by Richard Erkhov.

Llama-3.1-8B-Pruned-4-Layers - GGUF

Model creator: https://huggingface.co/Na0s/
Original model: https://huggingface.co/Na0s/Llama-3.1-8B-Pruned-4-Layers/

Name	Quant method	Size
Llama-3.1-8B-Pruned-4-Layers.Q2_K.gguf	Q2_K	2.59GB
Llama-3.1-8B-Pruned-4-Layers.IQ3_XS.gguf	IQ3_XS	2.86GB
Llama-3.1-8B-Pruned-4-Layers.IQ3_S.gguf	IQ3_S	2.99GB
Llama-3.1-8B-Pruned-4-Layers.Q3_K_S.gguf	Q3_K_S	2.98GB
Llama-3.1-8B-Pruned-4-Layers.IQ3_M.gguf	IQ3_M	3.07GB
Llama-3.1-8B-Pruned-4-Layers.Q3_K.gguf	Q3_K	3.25GB
Llama-3.1-8B-Pruned-4-Layers.Q3_K_M.gguf	Q3_K_M	3.25GB
Llama-3.1-8B-Pruned-4-Layers.Q3_K_L.gguf	Q3_K_L	3.49GB
Llama-3.1-8B-Pruned-4-Layers.IQ4_XS.gguf	IQ4_XS	3.63GB
Llama-3.1-8B-Pruned-4-Layers.Q4_0.gguf	Q4_0	1.7GB
Llama-3.1-8B-Pruned-4-Layers.IQ4_NL.gguf	IQ4_NL	3.8GB
Llama-3.1-8B-Pruned-4-Layers.Q4_K_S.gguf	Q4_K_S	3.79GB
Llama-3.1-8B-Pruned-4-Layers.Q4_K.gguf	Q4_K	3.97GB
Llama-3.1-8B-Pruned-4-Layers.Q4_K_M.gguf	Q4_K_M	3.97GB
Llama-3.1-8B-Pruned-4-Layers.Q4_1.gguf	Q4_1	4.14GB
Llama-3.1-8B-Pruned-4-Layers.Q5_0.gguf	Q5_0	4.52GB
Llama-3.1-8B-Pruned-4-Layers.Q5_K_S.gguf	Q5_K_S	4.52GB
Llama-3.1-8B-Pruned-4-Layers.Q5_K.gguf	Q5_K	4.62GB
Llama-3.1-8B-Pruned-4-Layers.Q5_K_M.gguf	Q5_K_M	4.62GB
Llama-3.1-8B-Pruned-4-Layers.Q5_1.gguf	Q5_1	4.89GB
Llama-3.1-8B-Pruned-4-Layers.Q6_K.gguf	Q6_K	5.31GB
Llama-3.1-8B-Pruned-4-Layers.Q8_0.gguf	Q8_0	6.87GB

Original model description:

library_name: transformers tags: - mergekit - merge pipeline_tag: text-generation

Na0s/Llama-3.1-8b-Pruned-4-Layers

This is a merge of meta-llama/Meta-Llama-3.1-8B created using mergekit, with respect to the paper "The Unreasonable Ineffectiveness of the Deeper Layers"

Merge Details

Merge Method

This model was merged using the passthrough merge method.

Models Merged

The following models were included in the merge:

meta-llama/Meta-Llama-3.1-8B

Configuration

The following YAML configuration was used to produce this model:

dtype: bfloat16
merge_method: passthrough
slices:
- sources:
  - layer_range: [0, 23]
    model: meta-llama/Meta-Llama-3.1-8B
- sources:
  - layer_range: [28, 32]
    model: meta-llama/Meta-Llama-3.1-8B

Evaluation

MMLU Pro 0-shot: 0.2642

Evaluation Data

[TIGER-AI-Lab/MMLU-Pro]

Environmental Impact

Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).