image/png

Qwen2.5-Gutenberg-Doppel-32B

Qwen/Qwen2.5-32B-Instruct finetuned on jondurbin/gutenberg-dpo-v0.1 and nbeerbower/gutenberg2-dpo.

Method

ORPO tuned with 2x A100 for 1.25 epochs.

Downloads last month
116
Safetensors
Model size
32.8B params
Tensor type
BF16
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for nbeerbower/Qwen2.5-Gutenberg-Doppel-32B

Base model

Qwen/Qwen2.5-32B
Finetuned
(158)
this model
Merges
12 models
Quantizations
13 models

Datasets used to train nbeerbower/Qwen2.5-Gutenberg-Doppel-32B