返回模型

说明文档

使用 ONNX 的 Wasm 运行时进行 SimVP 视频预测。你可以使用下面链接的 notebook 来训练自己的视频预测模型，该模型可以在浏览器中实时运行。考虑到其运行速度，SimVP 令人印象深刻 —— 详见论文中与其他方法的对比结果。

演示：

注意，这些示例可能训练不足。它们只是概念验证，用于确保整个流程（从训练到浏览器推理）正常工作。

多种模型（彩色）：https://josephrocca.github.io/SimVP-web/demo/simple.html
与上述相同，但为无限/递归生成：https://josephrocca.github.io/SimVP-web/demo/simple-infinite.html
移动 MNIST（灰度）：https://josephrocca.github.io/SimVP-web/demo/mmnist.html

训练与转换：

移动 MNIST：https://colab.research.google.com/gist/josephrocca/80c4a9814621d228dfb3d262a26a0806
使用自定义视频数据集进行训练：https://colab.research.google.com/gist/josephrocca/7be01186b14e07b5d54e80c245a35570

预训练与转换后的模型：

https://huggingface.co/rocca/simvp-web

扩散实验：

rocca/simvp-web

作者 rocca

↓ 0 ♥ 0

创建时间: 2022-12-01 05:00:27+00:00

更新时间: 2023-02-17 04:51:16+00:00

在 Hugging Face 上查看

文件 (48)

.gitattributes

README.md

onnx/fortnite/README.md

onnx/fortnite/fortnite_epoch13_1.7mil-steps_quantized.onnx ONNX

onnx/fortnite/fortnite_epoch8_1mil-steps_quantized.onnx ONNX

onnx/fortnite/input_output_data.json.br

onnx/fortnite_diffusion/README.md

onnx/fortnite_diffusion/fortnite_diffusion_epoch1_129k-steps.onnx ONNX

onnx/fortnite_diffusion/fortnite_diffusion_epoch1_129k-steps_quantized.onnx ONNX

onnx/fortnite_diffusion/fortnite_diffusion_epoch2_258k-steps.onnx ONNX

onnx/fortnite_diffusion/fortnite_diffusion_epoch2_258k-steps_quantized.onnx ONNX

onnx/fortnite_diffusion/input_output_data.json.br

onnx/fortnite_diffusion_2/README.md

onnx/fortnite_diffusion_2/fortnite_diffusion_2_epoch1_129k-steps.onnx ONNX

onnx/fortnite_diffusion_2/fortnite_diffusion_2_epoch1_129k-steps_quantized.onnx ONNX

onnx/fortnite_diffusion_2/fortnite_diffusion_2_epoch4_516k-steps.onnx ONNX

onnx/fortnite_diffusion_2/fortnite_diffusion_2_epoch4_516k-steps_quantized.onnx ONNX

onnx/fortnite_diffusion_2/fortnite_diffusion_2_epoch8_1mil-steps.onnx ONNX

onnx/fortnite_diffusion_2/fortnite_diffusion_2_epoch8_1mil-steps_quantized.onnx ONNX

onnx/fortnite_diffusion_2/input_output_data.json.br

onnx/moving_mnist/README.md

onnx/moving_mnist/epoch-137-quantized.onnx ONNX

onnx/moving_mnist/epoch-137.onnx ONNX

onnx/moving_mnist/input_output_data.json

onnx/pokemon_gold/README.md

onnx/pokemon_gold/input_output_data.json.br

onnx/pokemon_gold/pokemon_gold_epoch10_1mil-steps_quantized.onnx ONNX

onnx/pokemon_gold/pokemon_gold_epoch15_1.5mil-steps_quantized.onnx ONNX

onnx/pokemon_gold_diffusion/README.md

onnx/pokemon_gold_diffusion/input_output_data.json.br

onnx/pokemon_gold_diffusion/pokemon_gold_diffusion_epoch1_100k-steps.onnx ONNX

onnx/pokemon_gold_diffusion/pokemon_gold_diffusion_epoch1_100k-steps_quantized.onnx ONNX

onnx/pokemon_gold_diffusion/pokemon_gold_diffusion_epoch2_200k-steps.onnx ONNX

onnx/pokemon_gold_diffusion/pokemon_gold_diffusion_epoch2_200k-steps_quantized.onnx ONNX

onnx/pokemon_gold_diffusion_2/README.md

onnx/pokemon_gold_diffusion_2/input_output_data.json.br

onnx/pokemon_gold_diffusion_2/pokemon_gold_diffusion_2_epoch11_1.1mil-steps_quantized.onnx ONNX

onnx/pokemon_gold_diffusion_2/pokemon_gold_diffusion_2_epoch1_100k-steps.onnx ONNX

onnx/pokemon_gold_diffusion_2/pokemon_gold_diffusion_2_epoch1_100k-steps_quantized.onnx ONNX

onnx/pokemon_gold_diffusion_2/pokemon_gold_diffusion_2_epoch5_500k-steps.onnx ONNX

onnx/pokemon_gold_diffusion_2/pokemon_gold_diffusion_2_epoch5_500k-steps_quantized.onnx ONNX

onnx/riseofgru/README.md

onnx/riseofgru/input_output_data.json

onnx/riseofgru/input_output_data.json.br

onnx/riseofgru/riseofgru_7238x2_steps_quantized.onnx ONNX

onnx/riseofgru/riseofgru_7238x4_steps.onnx ONNX

onnx/riseofgru/riseofgru_7238x4_steps_quantized.onnx ONNX

onnx/riseofgru/riseofgru_7238x5_steps_quantized.onnx ONNX