Downloads · 30 days
15
5% of all-time downloads
TiME-de-m
dschulmeist/TiME-de-m
TiME-de-m is a feature extraction model from dschulmeist. Use it when you need embeddings to search or compare text. It is set up for transformers. The card lists the license as apache-2.0.
Monolingual BERT-style encoder that outputs embeddings for German. Distilled from FacebookAI/xlm-roberta-large.
All-time downloads
323
Public
Parameters
236M
1.9 GB on disk
Likes
1
Public
Repo makeup
Click a slice to open those files.
.bin942 MB · 49%
At a glance
- Task
- Feature Extraction
- Library
- transformers
- License
- apache-2.0
- Model type
- bert
- Access
- Public
- Created
- Aug 25, 2025
- Updated
- Nov 16, 2025
- SHA
- 314bf180
Datasets
Model card
From the Hugging Face model README
TiME German (de, m)
Monolingual BERT-style encoder that outputs embeddings for German. Distilled from FacebookAI/xlm-roberta-large.
Specs
- language: German (de)
- size: m
- architecture: BERT encoder
- layers: 6
- hidden size: 768
- intermediate size: 3072
Usage (mean pooled embeddings)
from transformers import AutoTokenizer, AutoModel
import torch
repo = "dschulmeist/TiME-de-m"
tok = AutoTokenizer.from_pretrained(repo)
mdl = AutoModel.from_pretrained(repo)
def mean_pool(last_hidden_state, attention_mask):
mask = attention_mask.unsqueeze(-1).type_as(last_hidden_state)
return (last_hidden_state * mask).sum(1) / mask.sum(1).clamp(min=1e-9)
inputs = tok(["example sentence"], padding=True, truncation=True, return_tensors="pt")
outputs = mdl(**inputs)
emb = mean_pool(outputs.last_hidden_state, inputs['attention_mask'])
print(emb.shape)
- Task
- Feature Extraction
- Library
- transformers
- Type
- bert
- License
- apache-2.0
- Languages
- de
- Created
- Aug 25, 2025
- Updated
- Nov 16, 2025