Hugging Peers

testmod_-_koen-llama2-13b-dpotrain_testver-gguf

☆ Save
by RichardErkhov
18.8K downloads 0 likes on Hugging Face updated

Model card

Quantization made by Richard Erkhov.

Github

Discord

Request more models

koen-llama2-13b-dpotrain_testver - GGUF

  • Model creator: https://huggingface.co/testmod/
  • Original model: https://huggingface.co/testmod/koen-llama2-13b-dpotrain_testver/
Name Quant method Size
koen-llama2-13b-dpotrain_testver.Q2_K.gguf Q2_K 4.52GB
koen-llama2-13b-dpotrain_testver.IQ3_XS.gguf IQ3_XS 4.99GB
koen-llama2-13b-dpotrain_testver.IQ3_S.gguf IQ3_S 5.27GB
koen-llama2-13b-dpotrain_testver.Q3_K_S.gguf Q3_K_S 5.27GB
koen-llama2-13b-dpotrain_testver.IQ3_M.gguf IQ3_M 5.57GB
koen-llama2-13b-dpotrain_testver.Q3_K.gguf Q3_K 1.25GB
koen-llama2-13b-dpotrain_testver.Q3_K_M.gguf Q3_K_M 1.96GB
koen-llama2-13b-dpotrain_testver.Q3_K_L.gguf Q3_K_L 0.3GB
koen-llama2-13b-dpotrain_testver.IQ4_XS.gguf IQ4_XS 0.0GB
koen-llama2-13b-dpotrain_testver.Q4_0.gguf Q4_0 0.0GB
koen-llama2-13b-dpotrain_testver.IQ4_NL.gguf IQ4_NL 0.0GB
koen-llama2-13b-dpotrain_testver.Q4_K_S.gguf Q4_K_S 0.0GB
koen-llama2-13b-dpotrain_testver.Q4_K.gguf Q4_K 0.0GB
koen-llama2-13b-dpotrain_testver.Q4_K_M.gguf Q4_K_M 0.0GB
koen-llama2-13b-dpotrain_testver.Q4_1.gguf Q4_1 0.0GB
koen-llama2-13b-dpotrain_testver.Q5_0.gguf Q5_0 0.0GB
koen-llama2-13b-dpotrain_testver.Q5_K_S.gguf Q5_K_S 0.0GB
koen-llama2-13b-dpotrain_testver.Q5_K.gguf Q5_K 0.0GB
koen-llama2-13b-dpotrain_testver.Q5_K_M.gguf Q5_K_M 0.0GB
koen-llama2-13b-dpotrain_testver.Q5_1.gguf Q5_1 0.0GB
koen-llama2-13b-dpotrain_testver.Q6_K.gguf Q6_K 0.0GB
koen-llama2-13b-dpotrain_testver.Q8_0.gguf Q8_0 0.0GB

Original model description:

language:

  • ko library_name: transformers pipeline_tag: text-generation license: apache-2.0

Input Models input text only.

Output Models generate text only.

Model Architecture
Auto-regressive language model based on the LLaMA2 transformer architecture.

Base Model
testmod/koen-llama2-13b-avg_testver

Training Objective
dpo trained with chosen/rejected pair datasets.