Instructions to use deepghs/timms with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- timm
How to use deepghs/timms with timm:
import timm model = timm.create_model("hf_hub:deepghs/timms", pretrained=True) - Notebooks
- Google Colab
- Kaggle
metadata
pipeline_tag: image-classification
base_model:
- timm/convformer_m36.sail_in22k_ft_in1k
- timm/convmixer_1024_20_ks9_p14.in1k
- timm/convnext_nano.r384_in12k_ft_in1k
- timm/convnext_pico.d1_in1k
- timm/convnextv2_atto.fcmae_ft_in1k
- timm/ecaresnet50t.a1_in1k
- timm/efficientnet_b5.sw_in12k
- timm/efficientvit_b1.r256_in1k
- timm/efficientvit_b2.r256_in1k
- timm/ghostnetv2_100.in1k
- timm/hgnetv2_b4.ssld_stage2_ft_in1k
- timm/mobilenetv2_100.ra_in1k
- timm/mobilevitv2_050.cvnets_in1k
- timm/nextvit_base.bd_in1k
- timm/regnetx_080.tv2_in1k
- timm/repvgg_b1.rvgg_in1k
- timm/resnetaa50.a1h_in1k
- timm/swin_base_patch4_window7_224.ms_in1k
- timm/vit_base_patch16_clip_384.laion2b_ft_in12k_in1k
- timm/vit_base_r50_s16_384.orig_in21k_ft_in1k
- timm/vit_little_patch16_reg4_gap_256.sbb_in1k
- timm/xcit_small_24_p16_224.fb_in1k
language:
- en
tags:
- timm
- image
- dghs-realutils
library_name: dghs-imgutils
ONNX export version from TIMM.
Models
22 models exported from TIMM in total.
ByobNet
2 models with model class ByobNet.
| Name | Params | Flops | Input Size | Can Classify | Features | Classes | Dataset | Model | Architecture | Created At |
|---|---|---|---|---|---|---|---|---|---|---|
| repvgg_b1.rvgg_in1k | 57.3M | 13.1G | 224 | True | 2048 | 1000 | imagenet-1k | ByobNet | repvgg_b1 | 2023-03-22 |
| mobilevitv2_050.cvnets_in1k | 1.4M | 464.6M | 256 | True | 256 | 1000 | imagenet-1k | ByobNet | mobilevitv2_050 | 2023-04-24 |
ConvMixer
1 model with model class ConvMixer.
| Name | Params | Flops | Input Size | Can Classify | Features | Classes | Dataset | Model | Architecture | Created At |
|---|---|---|---|---|---|---|---|---|---|---|
| convmixer_1024_20_ks9_p14.in1k | 24.4M | 6.0G | 224 | True | 1024 | 1000 | imagenet-1k | ConvMixer | convmixer_1024_20_ks9_p14 | 2023-04-24 |
ConvNeXt
3 models with model class ConvNeXt.
| Name | Params | Flops | Input Size | Can Classify | Features | Classes | Dataset | Model | Architecture | Created At |
|---|---|---|---|---|---|---|---|---|---|---|
| convnext_nano.r384_in12k_ft_in1k | 15.6M | 7.2G | 384 | True | 640 | 1000 | imagenet-1k | ConvNeXt | convnext_nano | 2024-12-31 |
| convnext_pico.d1_in1k | 9.0M | 1.4G | 224 | True | 512 | 1000 | imagenet-1k | ConvNeXt | convnext_pico | 2022-12-13 |
| convnextv2_atto.fcmae_ft_in1k | 3.7M | 547.3M | 224 | True | 320 | 1000 | imagenet-1k | ConvNeXt | convnextv2_atto | 2023-01-05 |
EfficientNet
2 models with model class EfficientNet.
| Name | Params | Flops | Input Size | Can Classify | Features | Classes | Dataset | Model | Architecture | Created At |
|---|---|---|---|---|---|---|---|---|---|---|
| efficientnet_b5.sw_in12k | 52.4M | 8.1G | 416 | True | 2048 | 11821 | imagenet-12k | EfficientNet | efficientnet_b5 | 2022-12-12 |
| mobilenetv2_100.ra_in1k | 3.5M | 300.8M | 224 | True | 1280 | 1000 | imagenet-1k | EfficientNet | mobilenetv2_100 | 2022-12-13 |
EfficientVit
2 models with model class EfficientVit.
| Name | Params | Flops | Input Size | Can Classify | Features | Classes | Dataset | Model | Architecture | Created At |
|---|---|---|---|---|---|---|---|---|---|---|
| efficientvit_b2.r256_in1k | 24.3M | 2.1G | 256 | True | 2560 | 1000 | imagenet-1k | EfficientVit | efficientvit_b2 | 2023-08-18 |
| efficientvit_b1.r256_in1k | 9.1M | 689.0M | 256 | True | 1600 | 1000 | imagenet-1k | EfficientVit | efficientvit_b1 | 2023-08-18 |
GhostNet
1 model with model class GhostNet.
| Name | Params | Flops | Input Size | Can Classify | Features | Classes | Dataset | Model | Architecture | Created At |
|---|---|---|---|---|---|---|---|---|---|---|
| ghostnetv2_100.in1k | 4.9M | 183.8M | 224 | True | 1280 | 1000 | imagenet-1k | GhostNet | ghostnetv2_100 | 2023-08-20 |
HighPerfGpuNet
1 model with model class HighPerfGpuNet.
| Name | Params | Flops | Input Size | Can Classify | Features | Classes | Dataset | Model | Architecture | Created At |
|---|---|---|---|---|---|---|---|---|---|---|
| hgnetv2_b4.ssld_stage2_ft_in1k | 19.8M | 2.8G | 224 | True | 2048 | 1000 | imagenet-1k | HighPerfGpuNet | hgnetv2_b4 | 2024-02-12 |
MetaFormer
1 model with model class MetaFormer.
| Name | Params | Flops | Input Size | Can Classify | Features | Classes | Dataset | Model | Architecture | Created At |
|---|---|---|---|---|---|---|---|---|---|---|
| convformer_m36.sail_in22k_ft_in1k | 57.0M | 12.8G | 224 | True | 576 | 1000 | imagenet-1k | MetaFormer | convformer_m36 | 2023-05-05 |
NextViT
1 model with model class NextViT.
| Name | Params | Flops | Input Size | Can Classify | Features | Classes | Dataset | Model | Architecture | Created At |
|---|---|---|---|---|---|---|---|---|---|---|
| nextvit_base.bd_in1k | 44.8M | 8.3G | 224 | True | 1024 | 1000 | imagenet-1k | NextViT | nextvit_base | 2024-02-11 |
RegNet
1 model with model class RegNet.
| Name | Params | Flops | Input Size | Can Classify | Features | Classes | Dataset | Model | Architecture | Created At |
|---|---|---|---|---|---|---|---|---|---|---|
| regnetx_080.tv2_in1k | 39.5M | 8.0G | 224 | True | 1920 | 1000 | imagenet-1k | RegNet | regnetx_080 | 2023-03-21 |
ResNet
2 models with model class ResNet.
| Name | Params | Flops | Input Size | Can Classify | Features | Classes | Dataset | Model | Architecture | Created At |
|---|---|---|---|---|---|---|---|---|---|---|
| resnetaa50.a1h_in1k | 25.6M | 5.2G | 224 | True | 2048 | 1000 | imagenet-1k | ResNet | resnetaa50 | 2023-04-05 |
| ecaresnet50t.a1_in1k | 25.6M | 4.3G | 224 | True | 2048 | 1000 | imagenet-1k | ResNet | ecaresnet50t | 2023-04-05 |
SwinTransformer
1 model with model class SwinTransformer.
| Name | Params | Flops | Input Size | Can Classify | Features | Classes | Dataset | Model | Architecture | Created At |
|---|---|---|---|---|---|---|---|---|---|---|
| swin_base_patch4_window7_224.ms_in1k | 87.7M | 15.2G | 224 | True | 1024 | 1000 | imagenet-1k | SwinTransformer | swin_base_patch4_window7_224 | 2023-03-18 |
VisionTransformer
3 models with model class VisionTransformer.
| Name | Params | Flops | Input Size | Can Classify | Features | Classes | Dataset | Model | Architecture | Created At |
|---|---|---|---|---|---|---|---|---|---|---|
| vit_base_r50_s16_384.orig_in21k_ft_in1k | 86.6M | 49.5G | 384 | True | 768 | 1000 | imagenet-1k | VisionTransformer | vit_base_r50_s16_384 | 2022-12-23 |
| vit_base_patch16_clip_384.laion2b_ft_in12k_in1k | 86.4M | 49.4G | 384 | True | 768 | 1000 | imagenet-1k | VisionTransformer | vit_base_patch16_clip_384 | 2022-11-11 |
| vit_little_patch16_reg4_gap_256.sbb_in1k | 22.4M | 5.7G | 256 | True | 320 | 1000 | imagenet-1k | VisionTransformer | vit_little_patch16_reg4_gap_256 | 2024-05-10 |
Xcit
1 model with model class Xcit.
| Name | Params | Flops | Input Size | Can Classify | Features | Classes | Dataset | Model | Architecture | Created At |
|---|---|---|---|---|---|---|---|---|---|---|
| xcit_small_24_p16_224.fb_in1k | 47.6M | 8.9G | 224 | True | 384 | 1000 | imagenet-1k | Xcit | xcit_small_24_p16_224 | 2023-04-13 |