YOLOv5/SparseML - Unable to find any modifiers in given recipe

34 Views Asked by At

I'm trying to train a YOLOv5s model with SparseML. (I don't know if this is important, but I am training in Google Colab). When I run train.py, I get this error:

ValueError: Unable to find any modifiers in given recipe. Modifiers must be listed as lists under yaml keys that include 'modifiers' in their name. Those keys and lists may also be nested under an extra key for staged recipes.

This is my recipe.yaml:

---
layout: null
title: ETS2 vehicle Detection Dataset Recipe
---

modifiers:
    - !EpochRangeModifier
        start_epoch: 0.0
        end_epoch: 250.0

    - !SetLearningRateModifier
        start_epoch: 5.0
        learning_rate: 0.1

    - !LearningRateModifier
        start_epoch: 0.0
        end_epoch: 25.0
        lr_class: MultiStepLR
        lr_kwargs:
            gamma: 0.9
            milestones: [2.0, 5.5, 10.0]
        init_lr: 0.1

    - !GMPruningModifier
        start_epoch: 50.0
        end_epoch: 100.0
        update_frequency: 1.0
        init_sparsity: 0.05
        final_sparsity: 0.65
        params: ['blocks.1.conv']
    
    - !QuantizationModifier
        start_epoch: 100.0

    - !TrainableParamsModifier
        params: ['blocks.1.conv']

    - !SetWeightDecayModifier
        start_epoch: 5.0
        weight_decay: 0.0
    
    - !ConstantPruningModifier
        params: ['blocks.1.conv]

What am I doing wrong with this, and how could I resolve this issue?

0

There are 0 best solutions below