Models available for E2E testing
DistilGPT-2 - Lightweight text generation
| Small Test | 7 tokens |
| Large Test | 128 tokens |
BERT base - Masked language model
| Small Test | 7 tokens |
| Large Test | 128 tokens |
RoBERTa base - Robust BERT variant
| Small Test | 7 tokens |
| Large Test | 128 tokens |
T5 small - Text-to-text transformer
| Small Test | 7 tokens |
| Large Test | 128 tokens |
DistilBERT - Smaller, faster BERT variant
| Small Test | 7 tokens |
| Large Test | 128 tokens |
ALBERT - Parameter-efficient BERT variant
| Small Test | 7 tokens |
| Large Test | 128 tokens |
Sentence-BERT - Text embeddings for semantic search
| Small Test | 16 tokens |
| Large Test | 128 tokens |
XLNet - Generalized autoregressive pretraining
| Small Test | 7 tokens |
| Large Test | 128 tokens |
ELECTRA - Efficient pre-training with replaced token detection
| Small Test | 7 tokens |
| Large Test | 128 tokens |
DeBERTa - Decoding-enhanced BERT with disentangled attention
| Small Test | 7 tokens |
| Large Test | 128 tokens |
DeBERTa v3 - Latest version with improved performance
| Small Test | 7 tokens |
| Large Test | 128 tokens |
MPNet - Masked and Permuted Pre-training
| Small Test | 7 tokens |
| Large Test | 128 tokens |
XLM-RoBERTa - Cross-lingual RoBERTa (100 languages)
| Small Test | 7 tokens |
| Large Test | 128 tokens |
DistilRoBERTa - Distilled RoBERTa for faster inference
| Small Test | 7 tokens |
| Large Test | 128 tokens |
SqueezeBERT - Mobile-optimized BERT variant
| Small Test | 7 tokens |
| Large Test | 128 tokens |
MiniLM - Compact model with deep self-attention distillation
| Small Test | 7 tokens |
| Large Test | 128 tokens |
BART base - Denoising autoencoder for pretraining
| Small Test | 7 tokens |
| Large Test | 128 tokens |
BGE Small - BAAI General Embedding (384-dim)
| Small Test | 16 tokens |
| Large Test | 512 tokens |
BGE Base - BAAI General Embedding (768-dim)
| Small Test | 16 tokens |
| Large Test | 512 tokens |
E5 Small - Text embeddings for retrieval (384-dim)
| Small Test | 16 tokens |
| Large Test | 512 tokens |
E5 Base - Text embeddings for retrieval (768-dim)
| Small Test | 16 tokens |
| Large Test | 512 tokens |
GTE Small - General Text Embeddings (384-dim)
| Small Test | 16 tokens |
| Large Test | 512 tokens |
GTE Base - General Text Embeddings (768-dim)
| Small Test | 16 tokens |
| Large Test | 512 tokens |
ResNet-50 - Image classification (1000 classes)
| Small Test | 64x64x3 |
| Large Test | 224x224x3 |
Vision Transformer (ViT) - Image classification
| Small Test | 64x64x3 |
| Large Test | 224x224x3 |
ConvNeXt Tiny - Modern CNN architecture
| Small Test | 64x64x3 |
| Large Test | 224x224x3 |
MobileNetV2 - Efficient mobile architecture
| Small Test | 64x64x3 |
| Large Test | 224x224x3 |
DeiT Small - Data-efficient Image Transformer
| Small Test | 64x64x3 |
| Large Test | 224x224x3 |
Swin Transformer - Shifted window attention
| Small Test | 64x64x3 |
| Large Test | 224x224x3 |
EfficientNet-B0 - Compound scaling CNN
| Small Test | 64x64x3 |
| Large Test | 224x224x3 |
RegNet - Designing Network Design Spaces
| Small Test | 64x64x3 |
| Large Test | 224x224x3 |
BEiT - BERT Pre-Training of Image Transformers
| Small Test | 64x64x3 |
| Large Test | 224x224x3 |
DINOv2 - Self-supervised vision transformer
| Small Test | 64x64x3 |
| Large Test | 224x224x3 |
PoolFormer - MetaFormer baseline with pooling
| Small Test | 64x64x3 |
| Large Test | 224x224x3 |
LeViT - Vision Transformer in ConvNet's Clothing
| Small Test | 64x64x3 |
| Large Test | 224x224x3 |
EfficientNet-B1 - Larger EfficientNet variant
| Small Test | 64x64x3 |
| Large Test | 240x240x3 |
EfficientNet-B2 - Medium EfficientNet variant
| Small Test | 64x64x3 |
| Large Test | 260x260x3 |
DenseNet-121 - Densely Connected Networks
| Small Test | 64x64x3 |
| Large Test | 224x224x3 |
ConvNeXt Small - Larger ConvNeXt variant
| Small Test | 64x64x3 |
| Large Test | 224x224x3 |
ViT Large - Large Vision Transformer
| Small Test | 64x64x3 |
| Large Test | 224x224x3 |
DETR - End-to-end object detection with transformer
| Small Test | 64x64x3 |
| Large Test | 800x600x3 |
SegFormer - Semantic segmentation
| Small Test | 64x64x3 |
| Large Test | 512x512x3 |
CLIP - Image-text matching and zero-shot classification
Text + Image input
Wav2Vec2 - Speech recognition
Large Language Models using GGUF format. Requires Core GGUF runtime plugin (llama.cpp).
TinyLlama 1.1B - Small but capable chat model
| Format | GGUF |
| Small Test | 32 tokens |
| Large Test | 256 tokens |
Microsoft Phi-2 - 2.7B parameter small language model
| Format | GGUF |
| Small Test | 64 tokens |
| Large Test | 256 tokens |
Qwen2 0.5B - Ultra-small instruction-tuned model (Alibaba)
| Format | GGUF |
| Small Test | 32 tokens |
| Large Test | 128 tokens |
Meta Llama 3.2 1B - Latest small model optimized for mobile
| Format | GGUF |
| Small Test | 32 tokens |
| Large Test | 256 tokens |
Meta Llama 3.2 3B - Excellent quality/size ratio
| Format | GGUF |
| Small Test | 64 tokens |
| Large Test | 256 tokens |
DeepSeek Coder 1.3B - Code generation specialist
| Format | GGUF |
| Small Test | 64 tokens |
| Large Test | 256 tokens |
DeepSeek LLM 7B Chat - High-quality open model
| Format | GGUF |
| Small Test | 64 tokens |
| Large Test | 256 tokens |
StableLM 2 Zephyr 1.6B - Stability AI's chat model
| Format | GGUF |
| Small Test | 32 tokens |
| Large Test | 128 tokens |
Gemma 2B - Google's lightweight open model
| Format | GGUF |
| Small Test | 64 tokens |
| Large Test | 256 tokens |
OpenChat 3.5 - High-quality chat model (7B)
| Format | GGUF |
| Small Test | 64 tokens |
| Large Test | 256 tokens |
Mistral 7B Instruct - Leading open 7B model
| Format | GGUF |
| Small Test | 64 tokens |
| Large Test | 512 tokens |
config/models.yamlenabled: true to include in testsmake config to verifymake test to testmy_model:
enabled: true
category: nlp
axon_id: "hf/my-org/my-model@latest"
description: "My awesome model"
input_type: text
small_input:
tokens: 7
large_input:
tokens: 128