Efficient Hyperparameter Tuning with Ray Tune and YOLO26
Hyperparameter tuning searches for the set of hyperparameters that gives the best model performance. It runs trials with different hyperparameter values and evaluates each trial's performance.
Accelerate Tuning with Ultralytics YOLO26 and Ray Tune#
Ultralytics YOLO26 integrates Ray Tune for hyperparameter tuning. With Ray Tune, you can use advanced search strategies, parallelism, and early stopping to speed up the tuning process.
Ray Tune#
Ray Tune is a hyperparameter tuning library designed for efficiency and flexibility. It supports various search strategies, parallelism, and early stopping strategies, and integrates with popular machine learning frameworks, including Ultralytics YOLO26.
Integration with Weights & Biases#
YOLO26 also supports optional integration with Weights & Biases for monitoring the tuning process.
Installation#
To install the required packages, run:
# Install and update Ultralytics and Ray Tune packages
pip install -U ultralytics "ray[tune]" optuna
# Optionally install and enable W&B for per-trial logging
pip install wandb
yolo settings wandb=TrueUsage#
from ultralytics import YOLO
# Load a YOLO26n model
model = YOLO("yolo26n.pt")
# Start tuning hyperparameters for YOLO26n training on the COCO8 dataset
result_grid = model.tune(data="coco8.yaml", use_ray=True)tune() Method Parameters#
The tune() method in YOLO26 provides an easy-to-use interface for hyperparameter tuning with Ray Tune. It accepts several arguments that allow you to customize the tuning process. Below is a detailed explanation of each parameter:
| Parameter | Type | Description | Default Value |
|---|---|---|---|
data | str or list | The dataset YAML to run the tuner on, specifying the training and validation data paths and other dataset-specific settings. Classification instead takes a dataset directory or a built-in dataset name (e.g., imagenet10). A list trains each trial on every dataset and reports the mean task metric. | |
space | dict, optional | A dictionary defining the hyperparameter search space for Ray Tune. Each key corresponds to a hyperparameter name, and the value specifies the range of values to explore during tuning. If not provided, YOLO26 uses a default search space with various hyperparameters. | |
grace_period | int, optional | The grace period in epochs for the ASHA scheduler in Ray Tune. The scheduler will not terminate any trial before this number of epochs (capped at epochs). Not used with search_alg="bohb". | 10 |
gpu_per_trial | int, optional | The number of GPUs Ray allocates to each trial. Set this explicitly to use GPU training; when omitted, trials receive no GPU resources. | None |
iterations | int, optional | The maximum number of trials to run during tuning. This parameter helps control the total number of hyperparameter combinations tested, ensuring the tuning process does not run indefinitely. | 300 |
search_alg | str or Searcher, optional | Ray Tune search strategy to use. You can pass a Ray searcher object directly, or use a supported string such as optuna, hyperopt, bayesopt, bohb, hebo, nevergrad, zoopt, ax, or random. String-based searchers automatically use the default task metric and mode="max". The default uses Optuna multivariate TPE with parallel-aware suggestions. | "optuna" |
**train_args | dict, optional | Additional arguments to pass to the train() method during tuning. These arguments can include settings like the number of training epochs, batch size, and other training-specific configurations. | {} |
By customizing these parameters, you can fine-tune the hyperparameter optimization process to suit your specific needs and available computational resources.
Default Search Space Description#
The following table lists the default search space parameters for hyperparameter tuning in YOLO26 with Ray Tune. Each parameter has a specific value range defined by tune.uniform(), except close_mosaic, which uses tune.randint().
| Parameter | Range | Description |
|---|---|---|
lr0 | tune.uniform(1e-5, 1e-2) | Initial learning rate that controls the step size during optimization. Higher values speed up training but may cause instability. |
lrf | tune.uniform(0.01, 1.0) | Final learning rate factor that determines how much the learning rate decreases by the end of training. |
momentum | tune.uniform(0.7, 0.98) | Momentum factor for the optimizer that helps accelerate training and overcome local minima. |
weight_decay | tune.uniform(0.0, 0.001) | Regularization parameter that prevents overfitting by penalizing large weight values. |
warmup_epochs | tune.uniform(0.0, 5.0) | Number of epochs with gradually increasing learning rate to stabilize early training. |
warmup_momentum | tune.uniform(0.0, 0.95) | Initial momentum value that gradually increases during the warmup period. |
box | tune.uniform(1.0, 20.0) | Weight for the bounding box loss component, balancing localization accuracy in the model. |
cls | tune.uniform(0.1, 4.0) | Weight for the classification loss component, balancing class prediction accuracy in the model. |
cls_pw | tune.uniform(0.0, 1.0) | Class weighting power, from disabled (0.0) to full inverse-frequency weighting (1.0). |
dfl | tune.uniform(0.4, 12.0) | Weight for the box-distance regression component, emphasizing precise bounding box localization. |
hsv_h | tune.uniform(0.0, 0.1) | Hue augmentation range that introduces color variability to help the model generalize. |
hsv_s | tune.uniform(0.0, 0.9) | Saturation augmentation range that varies color intensity to improve robustness. |
hsv_v | tune.uniform(0.0, 0.9) | Value (brightness) augmentation range that helps the model perform under various lighting conditions. |
degrees | tune.uniform(0.0, 45.0) | Rotation augmentation range in degrees, improving recognition of rotated objects. |
translate | tune.uniform(0.0, 0.9) | Translation augmentation range that shifts images horizontally and vertically. |
scale | tune.uniform(0.0, 0.95) | Scaling augmentation range that simulates objects at different distances. |
shear | tune.uniform(0.0, 10.0) | Shear augmentation range in degrees, simulating perspective shifts. |
perspective | tune.uniform(0.0, 0.001) | Perspective augmentation range that simulates 3D viewpoint changes. |
flipud | tune.uniform(0.0, 1.0) | Vertical flip augmentation probability, increasing dataset diversity. |
fliplr | tune.uniform(0.0, 1.0) | Horizontal flip augmentation probability, useful for symmetrical objects. |
bgr | tune.uniform(0.0, 1.0) | BGR channel swap augmentation probability, helping with color invariance. |
mosaic | tune.uniform(0.0, 1.0) | Mosaic augmentation probability that combines four images into one training sample. |
mixup | tune.uniform(0.0, 1.0) | Mixup augmentation probability that blends two images and their labels together. |
cutmix | tune.uniform(0.0, 1.0) | Cutmix augmentation probability that combines image regions while maintaining local features. |
copy_paste | tune.uniform(0.0, 1.0) | In flip mode the fraction of eligible objects copied into an image, in mixup mode also the probability of running it. |
close_mosaic | tune.randint(0, 11) | Disables mosaic in the last N epochs to stabilize training before completion. |
Custom Search Space Example#
This example shows how to use a custom search space for hyperparameter tuning with Ray Tune and YOLO26. By providing a custom search space, you can focus the tuning process on specific hyperparameters of interest.
from ray import tune
from ultralytics import YOLO
# Define a YOLO model
model = YOLO("yolo26n.pt")
# Run Ray Tune on the model
result_grid = model.tune(
data="coco8.yaml",
space={"lr0": tune.uniform(1e-5, 1e-2)},
epochs=50,
use_ray=True,
)The code snippet above creates a YOLO model from the pretrained YOLO26n weights, then calls the tune() method with the dataset configuration coco8.yaml. It provides a custom search space for the initial learning rate lr0 as a dictionary with the key "lr0" and the value tune.uniform(1e-5, 1e-2). Finally, it passes additional training arguments, such as the number of epochs, directly to the tune method as epochs=50.
Search Algorithm Example#
You can select a Ray Tune search strategy by name. For string-based search algorithms, Ultralytics will automatically use the current task's default optimization metric and mode="max". If search_alg is not provided, Ultralytics uses Optuna multivariate TPE with parallel-aware suggestions.
from ultralytics import YOLO
# Define a YOLO model
model = YOLO("yolo26n.pt")
result_grid = model.tune(
data="coco8.yaml",
epochs=50,
search_alg="optuna",
use_ray=True,
)If you need to customize the searcher's sampler or override its metric and optimization mode, pass an instantiated Ray Tune searcher object instead of a string.
from ray.tune.search.optuna import OptunaSearch
from ultralytics import YOLO
# Define a YOLO model
model = YOLO("yolo26n.pt")
# Define a searcher
searcher = OptunaSearch(
metric="metrics/mAP50-95(B)",
mode="max",
)
result_grid = model.tune(
data="coco8.yaml",
epochs=50,
search_alg=searcher,
use_ray=True,
)Resuming An Interrupted Hyperparameter Tuning Session With Ray Tune#
You can resume an interrupted Ray Tune session by passing resume=True. You can optionally pass the directory name used by Ray Tune under runs/{task} to resume. Otherwise, it resumes the last interrupted session. You don't need to provide the iterations and space again, but you need to provide the rest of the training arguments again including data and epochs.
from ultralytics import YOLO
# Define a YOLO model
model = YOLO("yolo26n.pt")
# Resume previous run
results = model.tune(use_ray=True, data="coco8.yaml", epochs=50, resume=True)
# Resume Ray Tune run with name 'tune_exp_2'
results = model.tune(use_ray=True, data="coco8.yaml", epochs=50, name="tune_exp_2", resume=True)Processing Ray Tune Results#
After running a hyperparameter tuning experiment with Ray Tune, you might want to perform various analyses on the obtained results. This guide will take you through common workflows for processing and analyzing these results.
Loading Tune Experiment Results from a Directory#
model.tune(use_ray=True) returns a Ray ResultGrid and saves the experiment under runs/{task}/tune, or under the name you passed. To load those results after the original script has exited, resume the session, which restores the saved trials and returns the same ResultGrid:
from ultralytics import YOLO
model = YOLO("yolo26n.pt")
result_grid = model.tune(data="coco8.yaml", epochs=50, use_ray=True, resume=True)Basic Experiment-Level Analysis#
Get an overview of how trials performed. You can quickly check if there were any errors during the trials.
if result_grid.errors:
print("One or more trials failed!")
else:
print("No errors!")Basic Trial-Level Analysis#
Access individual trial hyperparameter configurations and the last reported metrics.
for i, result in enumerate(result_grid):
print(f"Trial #{i}: Configuration: {result.config}, Last Reported Metrics: {result.metrics}")Plotting the Entire History of Reported Metrics for a Trial#
You can plot the history of reported metrics for each trial to see how the metrics evolved over time.
import matplotlib.pyplot as plt
for i, result in enumerate(result_grid):
plt.plot(
result.metrics_dataframe["training_iteration"],
result.metrics_dataframe["metrics/mAP50-95(B)"],
label=f"Trial {i}",
)
plt.xlabel("Training Iterations")
plt.ylabel("mAP50-95")
plt.legend()
plt.show()Summary#
In this guide, we covered common workflows to analyze the results of experiments run with Ray Tune using Ultralytics. The key steps include loading the experiment results from a directory, performing basic experiment-level and trial-level analysis, and plotting metrics.
Explore further by looking into Ray Tune's Analyze Results docs page to get the most out of your hyperparameter tuning experiments.
FAQ#
To tune the hyperparameters of your Ultralytics YOLO26 model using Ray Tune, follow these steps:
-
Install the required packages:
pip install -U ultralytics "ray[tune]" optuna pip install wandb # optional for logging yolo settings wandb=True -
Load your YOLO26 model and start tuning:
from ultralytics import YOLO # Load a YOLO26 model model = YOLO("yolo26n.pt") # Start tuning with the COCO8 dataset result_grid = model.tune(data="coco8.yaml", use_ray=True)
This uses Ray Tune search strategies and parallelism to optimize your model's hyperparameters efficiently.
-
Ultralytics YOLO26 uses the following default hyperparameters for tuning with Ray Tune:
Parameter Value Range Description lr0tune.uniform(1e-5, 1e-2)Initial learning rate lrftune.uniform(0.01, 1.0)Final learning rate factor momentumtune.uniform(0.7, 0.98)Momentum weight_decaytune.uniform(0.0, 0.001)Weight decay warmup_epochstune.uniform(0.0, 5.0)Warmup epochs boxtune.uniform(1.0, 20.0)Box loss weight clstune.uniform(0.1, 4.0)Class loss weight cls_pwtune.uniform(0.0, 1.0)Class weighting power dfltune.uniform(0.4, 12.0)Box-distance loss weight hsv_htune.uniform(0.0, 0.1)Hue augmentation range translatetune.uniform(0.0, 0.9)Translation augmentation range These hyperparameters can be customized to suit your specific needs. For a complete list and more details, refer to the Hyperparameter Tuning guide.
To integrate Weights & Biases (W&B) with your Ultralytics YOLO26 tuning process:
-
Install W&B:
pip install wandb wandb login yolo settings wandb=True -
Run the tuner:
from ultralytics import YOLO # Load YOLO model model = YOLO("yolo26n.pt") # Each Ray trial creates its own W&B run in this project result_grid = model.tune(data="coco8.yaml", project="YOLO-Tuning", use_ray=True)
This setup will allow you to monitor the tuning process, track hyperparameter configurations, and visualize results in W&B.
-
Ray Tune offers numerous advantages for hyperparameter optimization:
- Advanced Search Strategies: Utilizes algorithms like Bayesian Optimization and HyperOpt for efficient parameter search.
- Parallelism: Supports parallel execution of multiple trials, significantly speeding up the tuning process.
- Early Stopping: Employs strategies like ASHA to terminate under-performing trials early, saving computational resources.
Ray Tune is built into Ultralytics YOLO26 through
model.tune(use_ray=True).To define a custom search space for your YOLO26 hyperparameter tuning with Ray Tune:
from ray import tune from ultralytics import YOLO model = YOLO("yolo26n.pt") search_space = {"lr0": tune.uniform(1e-5, 1e-2), "momentum": tune.uniform(0.7, 0.98)} result_grid = model.tune(data="coco8.yaml", space=search_space, use_ray=True)This customizes the range of hyperparameters like initial learning rate and momentum to be explored during the tuning process. For advanced configurations, refer to the Custom Search Space Example section.
Comments40
Ask a question, share your results, or help someone else.
supriamir
Hi, I finish the hyperparameter tune, how to use the best parameter in best_hyperparameters.ymal for training?
Currently i manually use the manually copy the htperparameter.
model.train( data=str(dataset_yaml), epochs=epochs, imgsz=imgsz, batch=batch, project=project, name=f"fold_{k}", workers=0, patience=100, exist_ok=True, cache=False, device=0,
lr0= 0.00254, lrf= 0.02718, momentum= 0.83225, weight_decay= 0.001, warmup_epochs= 2.22741, warmup_momentum= 0.72834, box= 8.86345, cls= 0.52365, dfl= 0.45493, hsv_h= 0.01544, hsv_s= 0.44438, hsv_v= 0.37867, degrees= 0.01115, translate= 0.10045, scale= 0.33157, shear= 0.0075, perspective= 6.0e-05, flipud= 0.00882, fliplr= 0.34916, bgr= 9.0e-05, mosaic= 0.63841, mixup= 0.00099, cutmix= 0.00361, copy_paste= 0.00051, close_mosaic= 6, )
Glenn JocherUltralyticsYes — you can load
best_hyperparameters.yamland unpack it directly intomodel.train()instead of copying values by hand; see the Ray Tune integration guide and feel free to keep follow-ups here in Discussions so the Ultralytics team can see them.from pathlib import Path import yaml from ultralytics import YOLO model = YOLO("yolo26n.pt") hyp_path = Path("runs/detect/tune/best_hyperparameters.yaml") with hyp_path.open() as f: hyp = yaml.safe_load(f) model.train( data=str(dataset_yaml), epochs=epochs, imgsz=imgsz, batch=batch, project=project, name=f"fold_{k}", workers=0, patience=100, exist_ok=True, cache=False, device=0, **hyp, )1
Emeyer87
Hello,
I've noticed that every Iteration with use_ray = True is downloading yolo11n.pt. I'm attempting to use this on an existing instance segmentation model I built but wanted to ensure the tuning is running against the model I've assigned, and not a fresh Detection model.
from ultralytics import YOLO from ray import tune
model_path = r"C:/Users/Me/Downloads/NashvilleSegmentation.pt" data_path = r"C:/Users/Me/Downloads/dataset/data.yaml" model=YOLO(model_path)
search_space = { "degrees": tune.uniform(0.0, 12.0) }
results = model.tune( data=data_path, epochs=20, # keep it tiny iterations=20, # exactly one trial optimizer="AdamW", space=search_space, device=0, seed=0, workers=2, amp=True, batch=16, patience=12, plots=False, save=False, project="test_tune", name="degrees_tune_01", lr0=0.00114, use_ray=True, gpu_per_trial=1
)
Emeyer87
Disregard...completely misunderstood what tuning was actually doing.
Glenn JocherUltralyticsGlad it’s sorted—quick tip: to tune your existing segmentation model, load your .pt first (model = YOLO("path/to/your.pt")) and then run model.tune(..., use_ray=True); Ray trials will reuse that same model/task—verify in the logs that task=segment.
Fatima-khanjar
Subject: Questions Regarding Using Ray Tune for YOLOv11 Hyperparameter Tuning
Hello!
I hope this message finds you well.
#hyperparameter tuning using tune() from ray import tune from ray.tune import ResultGrid from ray.tune.examples.mnist_pytorch import train_mnist from ultralytics import YOLO import os import torch import matplotlib.pyplot as plt # Define a YOLO model model = YOLO("yolo11s.pt").to('cuda') tune_dir = 'ray_results' os.makedirs(tune_dir, exist_ok=True) # Tune hyperparameters using ray tune result_grid = model.tune( data= "data.yaml", epochs=10, iterations=10, batch = 16, device = 0, workers=0, use_ray=True, gpu_per_trial =1, name = "tuning", project = tune_dir ) Now, load the results after the tuning experiment is finished exp_name = "tuning" experiment_path = f"{tune_dir}/{exp_name}" # Get the name of the experiment print(f"Loading results from {experiment_path}...") restored_tuner = tune.Tuner.restore(experiment_path, trainable=train_mnist) result_grid = restored_tuner.get_results() # Basic Experiment-Level Analysis: Check for errors if result_grid.errors: print("One or more trials failed!") else: print("No errors!") # Basic Trial-Level Analysis: Print configurations and last reported metrics for each trial for i, result in enumerate(result_grid): print(f"Trial #{i}: Configuration: {result.config}, Last Reported Metrics: {result.metrics}") # Plotting the Entire History of Reported Metrics for a Trial # Assuming "mean_accuracy" is part of the metrics for i, result in enumerate(result_grid): plt.plot( result.metrics_dataframe["training_iteration"], result.metrics_dataframe["mean_accuracy"], # Modify this depending on your metrics label=f"Trial {i}", ) plt.xlabel("Training Iterations") plt.ylabel("Mean Accuracy") plt.legend() plt.show()
#hyperparameter tuning using tuner.fit() import os from ray import tune from ray.tune import ResultGrid from ray.tune.examples.mnist_pytorch import train_mnist from ray.tune import Result from pathlib import Path tune_dir = Path("ray_results").resolve().as_uri() exp_name = "tune_analyzing_results" tuner = tune.Tuner( trainable=train_mnist, param_space={ "lr": tune.uniform(0.001, 0.1), "momentum": tune.uniform(0.1, 0.9), }, run_config=tune.RunConfig( name=exp_name, stop={"training_iteration": 100}, checkpoint_config=tune.CheckpointConfig( checkpoint_score_attribute="mean_accuracy", num_to_keep = 5, ), storage_path=tune_dir, ), tune_config=tune.TuneConfig(mode="max", metric="mean_accuracy", num_samples=3), ) result_grid: ResultGrid = tuner.fit() experiment_path = os.path.join(tune_dir, exp_name) print(f"Loading results from {experiment_path}...") restored_tuner = tune.Tuner.restore(experiment_path, trainable=train_mnist) result_grid = restored_tuner.get_results() # Check if there have been errors if result_grid.errors: print("One of the trials failed!") else: print("No errors!") num_results = len(result_grid) print("Number of results:", num_results) # Iterate over results for i, result in enumerate(result_grid): if result.error: print(f"Trial #{i} had an error:", result.error) continue print( f"Trial #{i} finished successfully with a mean accuracy metric of:", result.metrics["mean_accuracy"] ) results_df = result_grid.get_dataframe() results_df[["training_iteration", "mean_accuracy"]] print("Shortest training time:", results_df["time_total_s"].min()) print("Longest training time:", results_df["time_total_s"].max()) best_result_df = result_grid.get_dataframe( filter_metric="mean_accuracy", filter_mode="max" ) best_result_df[["training_iteration", "mean_accuracy"]] # Get the result with the maximum test set `mean_accuracy` best_result: Result = result_grid.get_best_result() # Get the result with the minimum `mean_accuracy` worst_performing_result: Result = result_grid.get_best_result( metric="mean_accuracy", mode="min" ) best_result.config best_result.path # Get the last Checkpoint associated with the best-performing trial best_result.checkpoint # Get the last reported set of metrics best_result.metrics result_df = best_result.metrics_dataframe result_df[["training_iteration", "mean_accuracy", "time_total_s"]] best_result.metrics_dataframe.plot("training_iteration", "mean_accuracy") ax = None for result in result_grid: label = f"lr={result.config['lr']:.3f}, momentum={result.config['momentum']}" if ax is None: ax = result.metrics_dataframe.plot("training_iteration", "mean_accuracy", label=label) else: result.metrics_dataframe.plot("training_iteration", "mean_accuracy", ax=ax, label=label) ax.set_title("Mean Accuracy vs. Training Iteration for All Trials") ax.set_ylabel("Mean Test Accuracy")
I have been following your guide on using Ray Tune for hyperparameter tuning with YOLOv11. I was able to run the experiment using the use_ray=True argument, and it worked successfully—thank you for the helpful explanation!
However, I still have a few questions and points of confusion:
-
Naming the Trials: How can I change or customize the names of the individual trials when using model.tune(..., use_ray=True)? Right now, the trials are auto-named, and I’d like more meaningful names to help keep track of different configurations.
-
Retrieving the Best Hyperparameters: After the tuning process is complete, I would like to extract the best hyperparameter configuration to reuse it for future training. I noticed in other examples (e.g., train_mnist) that you use tuner.fit() to retrieve and analyze results, including the best configuration. So my question is: when using model.tune() directly (as in the case of YOLOv11), is there a method to extract the best hyperparameters without using tuner.fit()? Or is it necessary to re-structure the experiment using tune.Tuner to get access to that kind of detailed analysis?
-
Using tune.Tuner() with YOLOv11: In your other examples, the tuning process is managed using tune.Tuner(...), but I’m not sure how to apply this approach to YOLOv11, since there’s no explicit trainable function like train_mnist. Should we always use model.tune() for YOLO models, or is there a way to integrate YOLOv11 into a Tuner workflow for more advanced customization and analysis?
Thank you so much for your time and support! I would really appreciate any clarification or guidance on these points.
Best regards,
Glenn JocherUltralytics@Fatima-khanjar Thank you for your detailed questions about Ray Tune integration with YOLO11!
-
Naming Trials: Currently, the high-level
model.tune()API doesn't provide direct control over individual trial naming. Ray Tune automatically generates these names. You can use thenameparameter to name the overall experiment (as you've done with "tuning"), but individual trial names remain auto-generated. -
Retrieving Best Hyperparameters: When using
model.tune()with Ray Tune, the best hyperparameters are automatically saved toruns/[task]/tune/best_hyperparameters.yaml. This file contains the optimal configuration and can be loaded for future use. You can also examinetune_results.csvfor all trial results. -
Using tune.Tuner(): While possible, creating a custom trainable function for YOLO with the low-level Ray Tune API requires more work. You'd need to wrap the YOLO training process in a function that configures and trains the model while reporting metrics to Ray Tune. For most users, the high-level
model.tune()API is recommended as it handles this integration automatically.
For more advanced customization, you might consider extending the Ray Tune integration in Ultralytics by referring to the source code in the
ultralytics/utils/tuner.pyfile.-
Fatima-khanjar
Hello! Thank you for your clarification. I will continue using model.tune() and use the following line to retrieve the best hyperparameters, since the best_hyperparameters.yaml file wasn’t automatically saved in the output directory of the Ray tuning:
best_config = analysis.get_best_config(metric="metrics/mAP50(B)",mode="max")`I have a question: I read that Optuna uses the TPE (Tree-structured Parzen Estimator) search algorithm, which is more powerful than the default search algorithms in Ray Tune. Since Ray Tune provides scalability and distributed tuning, is it possible to combine both—using Optuna’s TPE algorithm with Ray Tune while still calling model.tune()? If so, how can I integrate them?
I also have another question that’s unrelated. I noticed that YOLOv11 comes in different versions, from n to xl. I understand that the choice depends on the dataset size and the task complexity. But what are the actual criteria used to determine the complexity of a task?
best regrads,
Glenn JocherUltralytics@Fatima-khanjar
Regarding using TPE with Ray Tune - yes, it's possible! Ray Tune already supports TPE through the HyperOpt integration. You can use it by installing the HyperOpt package and specifying the search algorithm:
pip install hyperopt from ray.tune.search.hyperopt import HyperOptSearch from ray.tune.search import ConcurrencyLimiter hyperopt_search = HyperOptSearch(metric="metrics/mAP50(B)", mode="max") limited_search = ConcurrencyLimiter(hyperopt_search, max_concurrent=4) # Then in your model.tune() call model.tune( data="data.yaml", search_alg=limited_search, use_ray=True, # other parameters )For your question about YOLO11 model size selection, the criteria typically include:
- Computational resources: Larger models (l, x, xl) require more GPU memory and processing power
- Dataset complexity: More complex datasets with small/numerous objects benefit from larger models
- Inference speed requirements: Smaller models (n, s) offer faster inference for real-time applications
- Accuracy needs: Larger models generally provide higher accuracy at the cost of speed
You can check the performance trade-offs in the Ultralytics documentation to make an informed decision based on your specific constraints.
1Fatima-khanjar(edited)
Hello, I hope this message finds you well.
I followed your instructions to integrate Ray Tune with the TPE algorithm using HyperOpt:
hyperopt_search = HyperOptSearch(metric="metrics/mAP50(B)", mode="max") limited_search = ConcurrencyLimiter(hyperopt_search, max_concurrent=1)Then, I used the following code to tune hyperparameters with Ray Tune:
tuning_results = model.tune( data="data.yaml", epochs=50, # Number of epochs for each tuning iteration iterations=100, # Total number of tuning iterations batch=16, device=0, workers=0, grace_period=30, search_alg=limited_search, use_ray=True, gpu_per_trial=1, project=tune_dir )However, I encountered a syntax error: SyntaxError: 'search_alg' is not a valid YOLO argument. Could you please advise me on how to fix this?
I have another question regarding the output directory of Ray Tune. It shows that inside the runs folder, there are subfolders named train (equal to the number of iterations) and a tune folder that contains subdirectories also equal to the number of iterations.
My question is: should I focus on the train directories? Why are they generated, and what's their purpose?
Also, if I want to monitor the overall progress of the fine-tuning process — for example, in genetic tuning, we usually look at a graph of the fitness score vs. iteration — where should I look for that kind of visualization or information?
best regards,
Glenn JocherUltralytics(edited)@Fatima-khanjar
Regarding your issues with HyperOptSearch and Ray Tune integration:
-
The
search_algargument is passed throughmodel.tune(use_ray=True). It defaults to"optuna"(TPE) and accepts either a string (such as"optuna","hyperopt","bohb"or"random") or a Ray Tune searcher instance like yourHyperOptSearch, so you do not need to modifyultralytics/utils/tuner.py. -
About the directory structure:
- The
traindirectories (train1, train2, etc.) contain individual training runs for each trial with different hyperparameter sets - The
tunedirectory contains the aggregated results of all trials and should be your primary focus
- The
-
For monitoring progress:
- Check
tune/tune_results.csvfor all trial results - Look for
tune/best_fitness.pngwhich shows fitness score progression over iterations tune/best_hyperparameters.yamlcontains the optimal configuration found
- Check
The plots in the
tunedirectory (especiallytune_scatter_plots.pngif available) provide valuable visualizations of the hyperparameter search space and how different parameters affected model performance.-
-
Fatima-khanjar
Hello!
I hope this message finds you well.I'm trying to use Ray Tune for hyperparameter optimization with the following code:
results = model.tune( data="data.yaml", epochs=30, iterations=150, batch=16, optimizer="AdamW", plots=False, save=False, val=False, device=0, workers=8, use_ray=True, project="ray_results" )However, I'm getting a path-related error that I can't figure out:
FileNotFoundError: [WinError 3] The system cannot find the path specified...It seems to be related to the length of the path Ray is trying to create. Do you know how I can fix this?
Also, I have a question. I came across Ray Tune while working with Genetic Algorithms, and I was wondering: would you recommend using Ray Tune or Genetic Algorithms for hyperparameter optimization? Are they similar in how they work, or are there major differences between them? What advantages does Ray Tune offer over Genetic Algorithms?
Thank you for your help!
Fatima-khanjar
Hello Community!!
Just following up on my own post — I managed to solve the issue on Windows 11.
Here’s what worked for me:
- Opened gpedit.msc as Administrator
- Navigated to: Computer Configuration > Administrative Templates > System > Filesystem
- Enabled the setting “Enable Win32 long paths"
After doing this, the problem was resolved.
Posting this in case it helps someone else!
Glenn JocherUltralytics@Fatima-khanjar Thank you for sharing your solution to the Windows path length issue! This is helpful information that could benefit other users experiencing similar problems.
Regarding your question about Ray Tune versus Genetic Algorithms: They're different approaches to hyperparameter optimization. Ray Tune offers several advantages including:
- Multiple search strategies beyond genetic algorithms (Bayesian optimization, grid search, random search)
- Built-in parallel execution that efficiently utilizes available resources
- Early stopping via ASHA scheduler to terminate underperforming trials
- Distributed computing capabilities for scaling across multiple machines
Ultralytics actually supports both approaches - the internal tuner uses genetic algorithms focusing on mutation, while the Ray Tune integration provides more advanced search strategies and better parallelization. Ray Tune is generally recommended for more complex search spaces and when you have multiple GPUs available.
1
wisrovi
como lo combino con MLFlow?
wisrovi
tambien necesito:
model.add_callback("on_train_start", self.on_train_start) model.add_callback("on_train_end", self.on_train_end)
pero dichas funciones no se disparan
wisrovi
agradezco mucho sus valiosas aportaciones, de verdad que es una funcionalidad increible y magnifica
Glenn JocherUltralytics@wisrovi to integrate MLFlow with Ultralytics YOLO hyperparameter tuning, you'll need to create a custom Ray Tune reporter that logs metrics to MLFlow. While we don't have native MLFlow integration like Weights & Biases, you can implement it by adding MLFlow logging callbacks to your tuning script. The Ray Tune documentation provides guidance on MLFlow integration that you can adapt for YOLO workflows.
NegroAmigo
Which versions of YOLO are supported by Ray Tune? I have project with YOLOv8, will it be suitable to use with Ray Tune?
Glenn JocherUltralytics@NegroAmigo yes, YOLOv8 fully supports Ray Tune integration for hyperparameter tuning. You can use the same
model.tune()method shown in our Ray Tune integration guide by simply loading your YOLOv8 model (YOLO('yolov8n.pt')). The tuning process works identically across supported YOLO versions. For implementation details, see the Ultralytics Ray Tune documentation.
alifazeli00
Why is there no mention of beta 2(Adamw) in ultralithic documents?
Glenn JocherUltralytics@alifazeli00 the documentation currently focuses on the default hyperparameters and their tuning ranges. If you're looking for specific details on beta2 for AdamW, it might not be explicitly covered. For more information on hyperparameter tuning, you can explore the Ray Tune documentation linked in the guide.
alifazeli00
thank you, It has been mentioned that momentum is essentially beta1. Is this the same for beta2?
Glenn JocherUltralyticsBeta2 is not explicitly mentioned in the documentation as it typically defaults to 0.999 in AdamW, which is a common standard setting. For specific adjustments, you may need to modify the optimizer settings directly in your training script.
1
sirtris
Hi, you mention that the
.tune()function integrates well with wandb. Are there any plans to make it work with mlflow as well? When I try to use it now no model metric, parameters or artifacts, etc. are logged. It would also be nice to have one hyperparameter sweep as a parent run and the individual runs as child runs of that parent run. Best,
Glenn JocherUltralytics@sirtris currently, there are no plans to integrate the
.tune()function with MLflow, but we appreciate your feedback and will consider it for future updates.sirtris
Could you point me to some resources so I could maybe look into it doing it myself?
Glenn JocherUltralyticsFor resources on integrating MLflow with Ray Tune, I recommend checking the Ray Tune documentation for any updates or community contributions. This might provide insights into logging metrics and managing runs with MLflow.
sachmatkris
unfortunately that is not the issue because the results are exactly the same for different learning rates - impossible.
On Fri, 27 Sept 2024 at 19:06, Glenn Jocher @.***> wrote:
@sachmatkris https://github.com/sachmatkris it seems like the issue might be with the search space configuration. Ensure that the hyperparameters have enough variability and that the dataset path is correctly set. If the problem persists, consider adjusting the search space or reviewing the dataset configuration. For more details, you can refer to the Ray Tune documentation https://docs.ultralytics.com/integrations/ray-tune/.
— Reply to this email directly, view it on GitHub https://github.com/orgs/ultralytics/discussions/16505#discussioncomment-10777716, or unsubscribe https://github.com/notifications/unsubscribe-auth/A3WH6HFTKYE2WLTJVGR6HWTZYV7A5AVCNFSM6AAAAABO5HD4QCVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTANZXG43TCNQ . You are receiving this because you were mentioned.Message ID: @.*** com>
Glenn JocherUltralytics@sachmatkris if you're seeing identical results for different learning rates, it might be worth checking the hyperparameter search space and ensuring the dataset configuration is correct. Adjusting these settings could help resolve the issue. For further guidance, please refer to the Ray Tune documentation.
sachmatkris
I found out the issue was setting device = 0 in the model.tune() and specifying absolute path for the .data.yaml, which is not the same what YOLO does when started from within an opened folder on VScode.
sachmatkris(edited)
@sachmatkris it seems like you've identified the issue with the device setting and path specification. If you need further assistance, feel free to ask!
Another issue I have is using the following call all of my experiments with ray_tune return the same metrics:
search_space = { 'lr0': tune.loguniform(1e-5, 1e-1), 'momentum': tune.uniform(0.6, 0.95), }result_grid = model.tune(data = 'C:/Users/Kristijonas/Desktop/Mokslo šaknys karčios/AI/Projects/masinu_numeriai/formated_dataset_new/data.yaml', imgsz = 1280, space = search_space, batch = 16, epochs = 50, patience = 5, use_ray = True, iterations = 20, device='cuda' )e.g.
Glenn JocherUltralytics@sachmatkris it seems like the issue might be with the search space configuration. Ensure that the hyperparameters have enough variability and that the dataset path is correctly set. If the problem persists, consider adjusting the search space or reviewing the dataset configuration. For more details, you can refer to the Ray Tune documentation.