arxiv-training

dmnv/arxiv-training

arxiv-training is a machine learning model from dmnv. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as apache-2.0.

should probably proofread and complete it, then remove this comment. --

Downloads · 30 days

3% of all-time downloads

All-time downloads

118

Public

Parameters

65.8M

263 MB on disk

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors263 MB · 100%

At a glance

License: apache-2.0
Model type: distilbert
Access: Public
Created: Apr 9, 2026
Updated: Apr 9, 2026
SHA: 952b057c

Base models

Model card

From the Hugging Face model README

arxiv-training

This model is a fine-tuned version of distilbert-base-cased on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 1.4004
Accuracy: 0.8076
F1: 0.8119

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
train_batch_size: 32
eval_batch_size: 8
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 0.1
num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
1.0381	1.0	1563	0.9201	0.7216	0.7432
0.784	2.0	3126	0.8248	0.7456	0.7639
0.6083	3.0	4689	0.8426	0.7552	0.7729
0.4241	4.0	6252	0.9093	0.7754	0.7897
0.2963	5.0	7815	1.0223	0.7934	0.8014
0.1993	6.0	9378	1.1686	0.7946	0.8027
0.1475	7.0	10941	1.2583	0.7984	0.8043
0.1046	8.0	12504	1.3767	0.8052	0.8084
0.0941	9.0	14067	1.4422	0.8082	0.8094
0.0781	10.0	15630	1.4518	0.8066	0.8079

Framework versions

Transformers 4.36.2
Pytorch 2.0.1+cu118
Datasets 4.8.4
Tokenizers 0.15.2

Type: distilbert
License: apache-2.0
Created: Apr 9, 2026
Updated: Apr 9, 2026