RichardErkhov/grimjim_-_Llama-3-Instruct-8B-SPPO-Iter3-SimPO-merge-gguf

Quantization made by Richard Erkhov.

Github

Discord

Request more models

Llama-3-Instruct-8B-SPPO-Iter3-SimPO-merge - GGUF

• Model creator: https://huggingface.co/grimjim/
• Original model: https://huggingface.co/grimjim/Llama-3-Instruct-8B-SPPO-Iter3-SimPO-merge/

Name	Quant method	Size
Llama-3-Instruct-8B-SPPO-Iter3-SimPO-merge.Q2_K.gguf	Q2_K	2.96GB
Llama-3-Instruct-8B-SPPO-Iter3-SimPO-merge.IQ3_XS.gguf	IQ3_XS	3.28GB
Llama-3-Instruct-8B-SPPO-Iter3-SimPO-merge.IQ3_S.gguf	IQ3_S	3.43GB
Llama-3-Instruct-8B-SPPO-Iter3-SimPO-merge.Q3_K_S.gguf	Q3_K_S	3.41GB
Llama-3-Instruct-8B-SPPO-Iter3-SimPO-merge.IQ3_M.gguf	IQ3_M	3.52GB
Llama-3-Instruct-8B-SPPO-Iter3-SimPO-merge.Q3_K.gguf	Q3_K	3.74GB
Llama-3-Instruct-8B-SPPO-Iter3-SimPO-merge.Q3_K_M.gguf	Q3_K_M	3.74GB
Llama-3-Instruct-8B-SPPO-Iter3-SimPO-merge.Q3_K_L.gguf	Q3_K_L	4.03GB
Llama-3-Instruct-8B-SPPO-Iter3-SimPO-merge.IQ4_XS.gguf	IQ4_XS	4.18GB
Llama-3-Instruct-8B-SPPO-Iter3-SimPO-merge.Q4_0.gguf	Q4_0	4.34GB
Llama-3-Instruct-8B-SPPO-Iter3-SimPO-merge.IQ4_NL.gguf	IQ4_NL	4.38GB
Llama-3-Instruct-8B-SPPO-Iter3-SimPO-merge.Q4_K_S.gguf	Q4_K_S	4.37GB
Llama-3-Instruct-8B-SPPO-Iter3-SimPO-merge.Q4_K.gguf	Q4_K	4.58GB
Llama-3-Instruct-8B-SPPO-Iter3-SimPO-merge.Q4_K_M.gguf	Q4_K_M	4.58GB
Llama-3-Instruct-8B-SPPO-Iter3-SimPO-merge.Q4_1.gguf	Q4_1	4.78GB
Llama-3-Instruct-8B-SPPO-Iter3-SimPO-merge.Q5_0.gguf	Q5_0	5.21GB
Llama-3-Instruct-8B-SPPO-Iter3-SimPO-merge.Q5_K_S.gguf	Q5_K_S	5.21GB
Llama-3-Instruct-8B-SPPO-Iter3-SimPO-merge.Q5_K.gguf	Q5_K	5.34GB
Llama-3-Instruct-8B-SPPO-Iter3-SimPO-merge.Q5_K_M.gguf	Q5_K_M	5.34GB
Llama-3-Instruct-8B-SPPO-Iter3-SimPO-merge.Q5_1.gguf	Q5_1	5.65GB
Llama-3-Instruct-8B-SPPO-Iter3-SimPO-merge.Q6_K.gguf	Q6_K	6.14GB
Llama-3-Instruct-8B-SPPO-Iter3-SimPO-merge.Q8_0.gguf	Q8_0	7.95GB

Original model description:

base_model: - princeton-nlp/Llama-3-Instruct-8B-SimPO - UCLA-AGI/Llama-3-Instruct-8B-SPPO-Iter3 library_name: transformers tags: - mergekit - merge license: llama3 pipeline_tag: text-generation

Llama-3-Instruct-8B-SPPO-Iter3-SimPO-merge

This is a merge of pre-trained language models created using mergekit.

Built with Meta Llama 3.

Merge Details

Merge Method

This model was merged using the SLERP merge method.

Models Merged

The following models were included in the merge:

• princeton-nlp/Llama-3-Instruct-8B-SimPO
• UCLA-AGI/Llama-3-Instruct-8B-SPPO-Iter3

Configuration

The following YAML configuration was used to produce this model:

slices:
- sources:
  - model: princeton-nlp/Llama-3-Instruct-8B-SimPO
    layer_range:
    - 0
    - 32
  - model: UCLA-AGI/Llama-3-Instruct-8B-SPPO-Iter3
    layer_range:
    - 0
    - 32
merge_method: slerp
base_model: UCLA-AGI/Llama-3-Instruct-8B-SPPO-Iter3
parameters:
  t:
  - filter: self_attn
    value:
    - 0
    - 0.5
    - 0.3
    - 0.7
    - 1
  - filter: mlp
    value:
    - 1
    - 0.5
    - 0.7
    - 0.3
    - 0
  - value: 0.5
dtype: bfloat16