CoolFace
Modelpublic

RichardErkhov/jkazdan_-_llama-refusal-gguf

sourceHugging Faceupdated 2y agoView on Hugging Face
0likes673downloads
Model Card

Quantization made by Richard Erkhov.

Github

Discord

Request more models

llama-refusal - GGUF

  • —Model creator: https://huggingface.co/jkazdan/
  • —Original model: https://huggingface.co/jkazdan/llama-refusal/
NameQuant methodSize
llama-refusal.Q2_K.ggufQ2_K1.27GB
llama-refusal.IQ3_XS.ggufIQ3_XS1.38GB
llama-refusal.IQ3_S.ggufIQ3_S1.44GB
llama-refusal.Q3_K_S.ggufQ3KS1.44GB
llama-refusal.IQ3_M.ggufIQ3_M1.49GB
llama-refusal.Q3_K.ggufQ3_K1.57GB
llama-refusal.Q3_K_M.ggufQ3KM1.57GB
llama-refusal.Q3_K_L.ggufQ3KL1.69GB
llama-refusal.IQ4_XS.ggufIQ4_XS1.71GB
llama-refusal.Q4_0.ggufQ4_01.79GB
llama-refusal.IQ4_NL.ggufIQ4_NL1.79GB
llama-refusal.Q4_K_S.ggufQ4KS1.8GB
llama-refusal.Q4_K.ggufQ4_K1.88GB
llama-refusal.Q4_K_M.ggufQ4KM1.88GB
llama-refusal.Q4_1.ggufQ4_11.95GB
llama-refusal.Q5_0.ggufQ5_02.11GB
llama-refusal.Q5_K_S.ggufQ5KS2.11GB
llama-refusal.Q5_K.ggufQ5_K2.16GB
llama-refusal.Q5_K_M.ggufQ5KM2.16GB
llama-refusal.Q5_1.ggufQ5_12.28GB
llama-refusal.Q6_K.ggufQ6_K2.46GB
llama-refusal.Q8_0.ggufQ8_03.19GB

Original model description: --- libraryname: transformers license: llama3.2 basemodel: meta-llama/Llama-3.2-3B-Instruct tags:

  • —trl
  • —sft
  • —generatedfromtrainer model-index:
  • —name: llama-refusal results: [] ---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

llama-refusal

This model is a fine-tuned version of meta-llama/Llama-3.2-3B-Instruct on an unknown dataset.

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

  • —learning_rate: 2e-05
  • —trainbatchsize: 4
  • —evalbatchsize: 4
  • —seed: 42
  • —gradientaccumulationsteps: 4
  • —totaltrainbatch_size: 16
  • —optimizer: Use OptimizerNames.ADAMWTORCHFUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
  • —lrschedulertype: linear
  • —lrschedulerwarmup_ratio: 0.05
  • —num_epochs: 1

Training results

Framework versions

  • —Transformers 4.46.3
  • —Pytorch 2.4.0+cu121
  • —Datasets 2.20.0
  • —Tokenizers 0.20.3