Hugging Face
Models
Datasets
Spaces
Posts
Docs
Enterprise
Pricing
Log In
Sign Up
DUAL-GPO
/
zephyr-7b-gpo-u4-i1
like
0
Follow
DUAL Group
2
PEFT
TensorBoard
Safetensors
HuggingFaceH4/ultrafeedback_binarized
mistral
alignment-handbook
Generated from Trainer
trl
dpo
License:
apache-2.0
Model card
Files
Files and versions
Metrics
Training metrics
Community
Train
Use this model
3adc41b
zephyr-7b-gpo-u4-i1
/
train_results.json
lole25
Model save
3adc41b
verified
10 months ago
raw
Copy download link
history
blame
Safe
195 Bytes
{
"epoch"
:
5.0
,
"train_loss"
:
0.3429992333650589
,
"train_runtime"
:
12841.5285
,
"train_samples"
:
61135
,
"train_samples_per_second"
:
0.389
,
"train_steps_per_second"
:
0.097
}