--- title: Google Cloud Deep Learning VM Image Reviews meta_title: 'Google Cloud Deep Learning VM Image Reviews 2026: Details, Pricing, & Features | G2' meta_description: Filter 19 reviews by the users' company size, role or industry to find out how Google Cloud Deep Learning VM Image works for a business like yours. aggregate_rating: rating_value: 4.3 review_count: 19 scale: '5' date_modified: '2026-07-30' parent_category: name: Deep Learning url: https://www.g2.com/categories/deep-learning --- # Google Cloud Deep Learning VM Image Reviews **Vendor:** Google **Category:** [Artificial Neural Network Software](https://www.g2.com/categories/artificial-neural-network) **Average Rating:** 4.3/5.0 **Total Reviews:** 19 ## About Google Cloud Deep Learning VM Image Deep Learning VM Image Preconfigured VMs for deep learning applications. ## Google Cloud Deep Learning VM Image Reviews ### 1. Deep Learning Win **Rating:** 5.0/5.0 stars **Reviewed by:** Nia S. | Project Manager/Team Lead, Small-Business (50 or fewer emp.) **Reviewed Date:** July 29, 2026 **What do you like best about Google Cloud Deep Learning VM Image?** I like how quick it is to get a deep learning environment up and running. Having the frameworks and GPU support preconfigured saves time, so I can focus on building and testing models instead of setting everything up manually. **What do you dislike about Google Cloud Deep Learning VM Image?** The initial setup is easy, but costs can add up if you leave GPU instances running. I've also run into occasional compatibility issues when updating frameworks, so a bit more flexibility around version management would help. **What problems is Google Cloud Deep Learning VM Image solving and how is that benefiting you?** It eliminates the time and effort of manually configuring deep learning environments. Since everything is preconfigured, I can start training and testing models much faster, which improves productivity and lets me spend more time on development instead of setup. ### 2. Easy Setup with Preinstalled AI Frameworks and Strong GPU Performance **Rating:** 4.5/5.0 stars **Reviewed by:** LOKESH G. | Engineer.SGB TCS-FS CORE BANKING,Production, Information Technology and Services, Enterprise (> 1000 emp.) **Reviewed Date:** July 27, 2026 **What do you like best about Google Cloud Deep Learning VM Image?** I like how easy it is to set up, with popular AI frameworks already installed, and how well it works with GPUs. It saves me time and makes it straightforward to start building and training AI models right away. **What do you dislike about Google Cloud Deep Learning VM Image?** The setup is straightforward and easy to get through, but GPU availability can be limited at times. Also, the cost can add up quickly, especially for long-running workloads. **What problems is Google Cloud Deep Learning VM Image solving and how is that benefiting you?** Google Cloud Deep Learning VM Image saves me the time and effort of setting up machine learning environments. Because the required frameworks and drivers are already installed, I can start training and testing models right away. This improves my productivity and helps speed up development. ### 3. Deep Learning VM Images: GPU-Ready, Compatible Stacks in Minutes **Rating:** 4.0/5.0 stars **Reviewed by:** Luca P. | Chief Operations Officer DEQUA Studio | Formerly CTO in MarTech, Marketing and Advertising, Mid-Market (51-1000 emp.) **Reviewed Date:** July 26, 2026 **What do you like best about Google Cloud Deep Learning VM Image?** The pitch is simple and it holds up in practice: a Compute Engine VM that boots with the framework, the CUDA stack, the NVIDIA driver, and the Python environment already installed and already compatible with each other. That last part is the one that matters. Anyone who has assembled a training machine by hand knows the real time sink is not installing PyTorch, it is discovering that the driver you installed does not match the CUDA toolkit that the framework wheel was compiled against, and unwinding that at 6pm. The Deep Learning VM images ship with those pieces tested together, and in the time I have used them I have not once had to debug a driver and toolkit mismatch on a fresh instance. That is the whole value proposition in one sentence, and it delivers. Provisioning itself takes minutes. I launch either from the Cloud Marketplace listing, where I pick the framework, the machine type, and the GPU from a form, or from gcloud when I am scripting it. The image family naming is worth learning because it encodes exactly what you get: framework, CUDA version, and OS in the family name, so I can pin a specific combination for reproducibility or point at the latest tag in the family when I just want current. For a team that needs everyone training against the same stack, referencing one image family in a shared launch script settles the environment question before it starts. What actually comes preinstalled is broader than the framework alone. On top of PyTorch or TensorFlow the image carries a full working Python data stack, the kind of baseline that every project needs and nobody enjoys assembling: - numpy, scipy, and pandas for the data handling layer - scikit-learn and scikit-image for classical ML and preprocessing - matplotlib for the quick sanity-check plots during training - OpenCV and Pillow for image pipelines - JupyterLab, running and reachable, for interactive work The JupyterLab piece deserves its own sentence. It is preconfigured on the instance, so the path from creating a VM to typing into a notebook against a live GPU is genuinely short. When I want to prototype a data loader or eyeball a batch of augmented samples before committing to a long run, I am in a notebook on the actual training hardware within a few minutes of deciding I need one. No SSH tunnel gymnastics, no installing a kernel by hand. GPU support is the area where the preconfiguration earns its keep most visibly. Attach a GPU at creation time and the image handles the driver, CUDA, cuDNN, and NCCL. On multi-GPU machines NCCL being present and matched to the CUDA version means distributed data parallel training works on the first attempt instead of after an afternoon of version archaeology. I have gone from no infrastructure to a multi-GPU PyTorch DDP run in under an hour, and most of that hour was my own code, not the machine. Because these are ordinary Compute Engine instances underneath, everything I already know about GCE applies. Snapshots for saving a configured state, persistent disks I can detach and reattach to a bigger machine when a job outgrows its instance, Spot provisioning for workloads that tolerate interruption, per-second billing so a two-hour experiment costs two hours. The image adds the ML layer without taking anything away from the platform layer, and that composability is what makes it usable for real work rather than demos. The update cadence is steady. New image versions track framework releases at a reasonable distance, and pinning a family version keeps older projects stable while new projects start on current builds. The images are also free in themselves. You pay for the compute, the disk, and the accelerator, and the preconfigured environment costs nothing on top, which makes the decision to use one over a bare Debian image very easy. The same environments also exist as Deep Learning Containers, and the pairing turns out to be more useful than it sounds. I prototype on a VM image, and when the workflow graduates to something scheduled or orchestrated, the matching container carries an equivalent stack into whatever runs it. Moving a training job from an interactive VM to an automated pipeline without rebuilding the environment from scratch removed a step I used to dread, because that rebuild was historically where subtle version drift crept in. Access to the machine is unremarkable in the best way. SSH from the console works in the browser with no key ceremony when I am away from my own laptop, gcloud compute ssh handles it from a terminal, and port forwarding to the JupyterLab instance is one flag. None of this is unique to these images, but it means the surrounding workflow has no rough edges of its own, and the two minutes saved per session compound over a project. One more thing I did not expect to care about: the base and HPC image variants without a framework. When I need a machine with the CUDA plumbing done but want to bring my own environment through conda or a container, the base image gives me the annoying bottom layers solved and stays out of the way above that. It is the version of the product for people who have opinions about their Python setup, and I appreciate that it exists. **What do you dislike about Google Cloud Deep Learning VM Image?** The costs are not a flaw of the image, but they are the context you live in. The image is free and the A100 or H100 attached to it very much is not, and because everything works immediately, it is easy to leave a GPU instance idle overnight out of habit. I schedule automatic shutdowns and lean on Spot instances for anything interruptible, and after adopting both the bill stopped surprising me. New users should set that discipline up on day one, because the platform will happily bill an idle accelerator at full rate. GPU quota is the friction I hit most often in practice. A fresh project starts with a GPU quota of zero in most regions, so the first launch attempt fails with a quota error and you file an increase request and wait. Approval has usually been quick for me, but it is an unadvertised speed bump directly in front of a product whose selling point is speed, and it confuses people the first time. My workaround is boring: request quota in two or three regions ahead of need, so capacity or approval delays in one region do not stall a deadline. The framework lineup has narrowed over time. The current image families center on PyTorch and the generic base and HPC images, and older framework variants age out of the update cadence rather than being maintained indefinitely. That is a defensible engineering choice, and it matches where the ecosystem went, but if your stack depends on an older framework build you will find yourself pinning an aging image family and inheriting its stale packages. Reading the release notes before committing a long-lived project to a specific family is worth the ten minutes. Documentation is serviceable rather than good. The core pages on choosing and creating an image are fine, but the material around the edges, upgrading in place, the exact contents of a given image version, what changed between releases, requires more digging than it should. More than once I have answered a question by SSHing into the instance and checking package versions directly instead of finding it written down. The image family naming convention helps, but a per-image manifest published somewhere obvious would help more. Last, small but real: the preinstalled environment is opinionated, and if your project needs different versions of the bundled libraries you end up creating a separate conda environment anyway, at which point some of the preinstallation is dead weight. The base image is the escape hatch here, and once I understood that split, picking the right variant per project made the issue mostly disappear. **What problems is Google Cloud Deep Learning VM Image solving and how is that benefiting you?** Standing up a training machine meant creating a bare VM, installing the NVIDIA driver, matching a CUDA toolkit to it, matching cuDNN to the toolkit, matching the framework wheel to all three, and then discovering some pair disagreed and starting the forensic work. On a good day that was an afternoon. On a bad day it was two, and the second day was the one where I questioned my career. The Deep Learning VM collapses that entire sequence into a form on the Marketplace page. The benefit is not just the hours back, it is that the hours come back at the start of a project, exactly when momentum is worth the most. Reproducibility across a team is the second problem it quietly solves. Before, every person's training environment drifted from everyone else's, and works on my machine arguments consumed real meeting time. Now the launch script names an image family and a version, and every instance anyone creates from it is identical down to the driver. When a result does not replicate, the environment is off the suspect list immediately, which shortens every one of those investigations. Scaling experiments up and down stopped being an infrastructure event. The old pattern was to size a machine for the biggest job you expected and live with the waste the rest of the time, because reconfiguring was painful enough to avoid. With the image, an environment exists on whatever machine shape I launch, so a cheap CPU instance handles the data preparation, a single-GPU machine handles the debugging runs, and the multi-GPU machine exists only for the hours of the real training job. The environment is no longer welded to a specific box, and the spend follows the actual workload instead of the worst case. It changed how I treat short-lived experiments. When creating a working environment cost half a day, I hoarded configured machines and hesitated before trying anything that needed a different setup. When it costs five minutes, instances become disposable. I spin one up to test a hypothesis, keep the results in a bucket and the code in git, and delete the machine without ceremony. More ideas get tested because the price of testing one collapsed, and that shows up in the quality of what eventually ships. Onboarding collaborators went from a document to a command. The before-state was a setup guide that was outdated within a month and a new team member burning their first days fighting it. Now they run the launch script, wait for the boot, and open JupyterLab on an environment identical to mine. First useful contribution moved from the end of week one to the first afternoon, and I no longer maintain the setup guide at all. Hardware evaluation became something I actually do instead of something I estimate. Choosing between a T4, an L4, and an A100 for a given model used to be guesswork informed by spec sheets, because benchmarking each option meant building each environment. Now the same image boots on any of them, so I run the identical training script on two or three accelerator types for an hour each and let the throughput per euro decide. Several times the cheaper card won, and I would never have found that out under the old setup cost, because nobody benchmarks options that take a day each to prepare. The notebook-on-real-hardware workflow closed a gap I used to paper over. Prototyping locally on a laptop meant the code met the actual GPU, the actual CUDA version, and the actual data path only at training time, which is the worst moment to find a problem. With JupyterLab preconfigured on the training instance itself, exploration happens on the same stack the long runs use, and the class of bugs that only appeared in the transition from laptop to cloud simply stopped appearing. Prototype and production environment are the same machine now, and a whole category of surprises went away with the difference. ### 4. Excellent Preconfigured Environment for Fast AI/ML Development on Google Cloud **Rating:** 4.0/5.0 stars **Reviewed by:** Aditya G. | Sales Account Manager, Mid-Market (51-1000 emp.) **Reviewed Date:** May 27, 2026 **What do you like best about Google Cloud Deep Learning VM Image?** What I like best about Google Cloud Deep Learning VM Images is how quickly they let you deploy a fully configured AI/ML environment without spending hours on setup and dependency management. Key advantages: Pre-installed frameworks like TensorFlow, PyTorch, JupyterLab, CUDA, and NVIDIA drivers GPU-ready environments optimized for AI workloads Faster experimentation and model training **What do you dislike about Google Cloud Deep Learning VM Image?** One challenge with Google Cloud Deep Learning VM Images is managing costs and optimizing resources, especially when running long workloads on high-end GPU instances. There are a few other limitations as well. There can be an initial learning curve for users who aren’t familiar with Google Cloud networking and IAM. GPU availability may also be limited in certain regions at times, which can make it harder to get the exact instance you need. Finally, the preconfigured environments sometimes include packages or versions that don’t align perfectly with specific project requirements, so some adjustments may be necessary. **What problems is Google Cloud Deep Learning VM Image solving and how is that benefiting you?** Benefits include: Faster deployment of AI/ML projects with minimal setup time Reduced infrastructure and configuration overhead for engineering teams Improved productivity for developers and data scientists Easier access to GPU-powered computing for model training and inference ### 5. Plug-and-Play Deep Learning VM Image, but Offline Reliability Needs Work **Rating:** 3.5/5.0 stars **Reviewed by:** Benjamin G. | Ammunition Specialist, Small-Business (50 or fewer emp.) **Reviewed Date:** July 07, 2026 **What do you like best about Google Cloud Deep Learning VM Image?** Google Cloud Deep Learning VM Image works great, no headache setup. Plug and Play. **What do you dislike about Google Cloud Deep Learning VM Image?** Doesn't work well offline and shutsdown quite often. **What problems is Google Cloud Deep Learning VM Image solving and how is that benefiting you?** Offline accessibility and faster GPU upgrades. ### 6. Powerful tool to help modernise your business **Rating:** 4.5/5.0 stars **Reviewed by:** Ryan C. | Operations manager, Small-Business (50 or fewer emp.) **Reviewed Date:** July 18, 2026 **What do you like best about Google Cloud Deep Learning VM Image?** We are able to process our organisation key performance data much more quickly now **What do you dislike about Google Cloud Deep Learning VM Image?** It may be a free start but it soon becomes very costly. **What problems is Google Cloud Deep Learning VM Image solving and how is that benefiting you?** With it being open to developers we have been able to get assistance to get the user interface more suited to us ### 7. Google Cloud Deep Learning VM Image Review **Rating:** 5.0/5.0 stars **Reviewed by:** Ramcharn H. | Senior Customer Service Officer, Mid-Market (51-1000 emp.) **Reviewed Date:** October 19, 2023 **What do you like best about Google Cloud Deep Learning VM Image?** You can quickly provision a VM with everything you need for your deep learning project on Google Cloud. Deep Learning VM Image makes it simple and quick to create a VM image containing all the most popular. **What do you dislike about Google Cloud Deep Learning VM Image?** "I found documentation to be a bit confusing as to where to search for reference and examples" **What problems is Google Cloud Deep Learning VM Image solving and how is that benefiting you?** Google after being the leading search engine provider dived into the cloud frenzy and it didn't disappoint. Its pricing is competitive , it has most of the features that other cloud providers have." ### 8. I developed and trained various deep learning models. **Rating:** 4.5/5.0 stars **Reviewed by:** Ali Burak . | Small-Business (50 or fewer emp.) **Reviewed Date:** October 28, 2023 **What do you like best about Google Cloud Deep Learning VM Image?** It is easy to set up and use. Thanks to its performance and scalability, I can train larger and more complex models. **What do you dislike about Google Cloud Deep Learning VM Image?** The cost is very high, it's easy to use, but there aren't many options to create a more customized environment for myself. **What problems is Google Cloud Deep Learning VM Image solving and how is that benefiting you?** I used it to develop a natural language processing model in a project. I introduced speech recognition models for translation. ### 9. Most recommended tool for Deep Learning **Rating:** 5.0/5.0 stars **Reviewed by:** Prashant P. | Business Development Manager, Information Technology and Services, Small-Business (50 or fewer emp.) **Reviewed Date:** October 18, 2023 **What do you like best about Google Cloud Deep Learning VM Image?** The Google Cloud Deep Learning is really a market game changer product in deep learning. Its really very easy to use and has good performance an security. It is cost efficient. **What do you dislike about Google Cloud Deep Learning VM Image?** There is nothing that can be disliked. Google Cloud Deep Learning VM Image is awesome. **What problems is Google Cloud Deep Learning VM Image solving and how is that benefiting you?** It has benefited me in the documentation. It has preconfigured environments that saved my time. I can easily scale up and down as per my requirements. Easy for team collaboration. Overall it has saved my time a lot. ### 10. A good investment DL VM **Rating:** 4.0/5.0 stars **Reviewed by:** Ramakant S. | Jr. Data Scientist, Small-Business (50 or fewer emp.) **Reviewed Date:** November 02, 2023 **What do you like best about Google Cloud Deep Learning VM Image?** Nice tool for data scientists/ML engineers. It has all the latest modules and is easy to install. Pre-installed common dependencies and driver support for GPU/TPU and CLI options make it awesome. It makes possible a few hours of work done in a few clicks. **What do you dislike about Google Cloud Deep Learning VM Image?** Nothing major as of now. faced an issue while restarting of collab notebook instance. **What problems is Google Cloud Deep Learning VM Image solving and how is that benefiting you?** It saves hours of debugging, makes faster deployment, and is easy to use. ### 11. Great Cloud Platform **Rating:** 4.5/5.0 stars **Reviewed by:** Verified User in Staffing and Recruiting | Mid-Market (51-1000 emp.) **Reviewed Date:** October 18, 2023 **What do you like best about Google Cloud Deep Learning VM Image?** its a best tool I have used because of easy access and the User Interface is very nice, anyone can use it easily without have any training or knowledge about it **What do you dislike about Google Cloud Deep Learning VM Image?** I didnt face anything in this tool which i would say that this is not good coz I have never gone through with those things **What problems is Google Cloud Deep Learning VM Image solving and how is that benefiting you?** Its a great AI framework. it makes easy and fast to exemplify a vm image and the best part is we can also easily add cloud GPU and cloud TPU and also it supports latest machine learning framework which is a great thing. ### 12. It has the tools that you want, the best combination ever **Rating:** 4.0/5.0 stars **Reviewed by:** Verified User in Information Technology and Services | Small-Business (50 or fewer emp.) **Reviewed Date:** October 17, 2023 **What do you like best about Google Cloud Deep Learning VM Image?** The Deep Learning VM Images come pre-installed with popular machine learning frameworks such as TensorFlow, PyTorch, and scikit-learn. This is a significant time-saver as setting up these environments can be complex and time-consuming. **What do you dislike about Google Cloud Deep Learning VM Image?** While you only pay for the resources you use, Google Cloud services can become expensive, especially when using powerful GPUs or TPUs. This may not be cost-effective for small businesses or individual developers. **What problems is Google Cloud Deep Learning VM Image solving and how is that benefiting you?** Setting up a deep learning environment from scratch can be complex and time-consuming. It involves installing the operating system, configuring drivers, and installing multiple libraries and dependencies. ### 13. Best image for Deep learning Project **Rating:** 5.0/5.0 stars **Reviewed by:** Sankalp A. | Mid-Market (51-1000 emp.) **Reviewed Date:** October 20, 2023 **What do you like best about Google Cloud Deep Learning VM Image?** Easy and Fast implementation of deep learning project on google cloud. It comes with pre-installed required softwares like PyTorch, Scikit-learn, Tensorflow and many more. **What do you dislike about Google Cloud Deep Learning VM Image?** Nothing to dislike about Google cloud deep learning vm image. **What problems is Google Cloud Deep Learning VM Image solving and how is that benefiting you?** Created deep learning project with the help of google cloud vm image without worrying about initial setup. ### 14. A deep learning virtual machine on Google Cloud and Effortless Deep Learning Setup . **Rating:** 5.0/5.0 stars **Reviewed by:** prajakta k. | DBA, Mid-Market (51-1000 emp.) **Reviewed Date:** October 20, 2023 **What do you like best about Google Cloud Deep Learning VM Image?** Google Cloud is easy to use.the best support for the customer. ease implementation. and we use this daily. This Google Cloud has so many features. **What do you dislike about Google Cloud Deep Learning VM Image?** There is no such dislike for Google Cloud. **What problems is Google Cloud Deep Learning VM Image solving and how is that benefiting you?** This helps me save time. and reducing technical issues with software installation. faster AI model development and experimentation. ### 15. Configurable and extensible but has a learning curve **Rating:** 3.0/5.0 stars **Reviewed by:** Daniel O. | TIP Contributor, Small-Business (50 or fewer emp.) **Reviewed Date:** October 19, 2023 **What do you like best about Google Cloud Deep Learning VM Image?** It has a lot of nobs & configurations to adapt to the user's needs regardless of how simple or how complex. **What do you dislike about Google Cloud Deep Learning VM Image?** The learning curve to use the softwre is quite steep. **What problems is Google Cloud Deep Learning VM Image solving and how is that benefiting you?** Managing the installation and configuration of ML libraries as well as built-in GCP integration. ### 16. Google Cloud Deep Learning VM Image Review **Rating:** 4.0/5.0 stars **Reviewed by:** Udit S. | Professional 2:-Application Delivery, Enterprise (> 1000 emp.) **Reviewed Date:** October 17, 2023 **What do you like best about Google Cloud Deep Learning VM Image?** It helps in bullding Deep Learning projects very fast on Google Cloud Platform. Its amazing. It has broad support, fast prototyping features which are good. **What do you dislike about Google Cloud Deep Learning VM Image?** The user interface of the platform can be improved. **What problems is Google Cloud Deep Learning VM Image solving and how is that benefiting you?** Deep Learning projects, prototyping, accessibility are the features which are beneficial. ### 17. Google cloud deep learning review **Rating:** 3.5/5.0 stars **Reviewed by:** hanne d. | learning builder, Mid-Market (51-1000 emp.) **Reviewed Date:** October 18, 2023 **What do you like best about Google Cloud Deep Learning VM Image?** This app corresponds with other Google applications. **What do you dislike about Google Cloud Deep Learning VM Image?** There is a learning curve to get to know this. **What problems is Google Cloud Deep Learning VM Image solving and how is that benefiting you?** This app corresponds with other Google apps which makes working fast. ### 18. Good product for user interface & ease **Rating:** 4.5/5.0 stars **Reviewed by:** Verified User in Staffing and Recruiting | Small-Business (50 or fewer emp.) **Reviewed Date:** November 07, 2023 **What do you like best about Google Cloud Deep Learning VM Image?** There is not a ton of setup or learning curve with this - its nice to use immediately. **What do you dislike about Google Cloud Deep Learning VM Image?** I didn't use everything this product is intended for and therefore did not run into any problems. **What problems is Google Cloud Deep Learning VM Image solving and how is that benefiting you?** We used it since it had many things come pre-installed and would be fast & easy. ### 19. GoogleVMImage **Rating:** 4.0/5.0 stars **Reviewed by:** Verified User in Education Management | Small-Business (50 or fewer emp.) **Reviewed Date:** October 18, 2023 **What do you like best about Google Cloud Deep Learning VM Image?** Saves time in setting up the environment as some frameworks are already installed **What do you dislike about Google Cloud Deep Learning VM Image?** We need to careful while using resources as itcan cause too much cost for the organization **What problems is Google Cloud Deep Learning VM Image solving and how is that benefiting you?** With some of the frameworks already pre installed it saves time while setting up any environment - [View Google Cloud Deep Learning VM Image pricing details and edition comparison](https://www.g2.com/products/google-cloud-deep-learning-vm-image/reviews?section=pricing&secure%5Bexpires_at%5D=2026-07-30+11%3A14%3A47+-0500&secure%5Bsession_id%5D=9fd062a8-bb24-4bb3-b364-c300c039ec20&secure%5Btoken%5D=143c29e66f7a806177c05ec342e5596b78edc65566fb13683f6f66eae60bb533&format=llm_user) ## Google Cloud Deep Learning VM Image Integrations - [SharePoint Integration](https://www.g2.com/products/sharepoint-integration/reviews) ## Google Cloud Deep Learning VM Image Features **Core Functionality - Artificial Neural Network** - Neural Network Training - Neural Network Testing - Model Evaluation - Compliance **Data Handling - Artificial Neural Network** - Data Integration - Data Preprocessing **Performance - Artificial Neural Network** - Model Optimization - Scalability **Usability - Artificial Neural Network** - User Interface - Documentation & Support - Customizability **Advanced Features - Artificial Neural Network** - Deep Learning Capabilities - Transfer Learning - Real-Time Processing - Automated Model Tuning - Visualization Tools **Agentic AI - Artificial Neural Network** - Autonomous Task Execution - Multi-step Planning - Cross-system Integration - Adaptive Learning - Natural Language Interaction - Proactive Assistance - Decision Making ## Top Google Cloud Deep Learning VM Image Alternatives - [Keras](https://www.g2.com/products/keras/reviews) - 4.6/5.0 (64 reviews) - [AIToolbox](https://www.g2.com/products/aitoolbox/reviews) - 4.4/5.0 (35 reviews) - [Microsoft Cognitive Toolkit (Formerly CNTK)](https://www.g2.com/products/microsoft-cognitive-toolkit-formerly-cntk/reviews) - 4.2/5.0 (22 reviews)