At the beginning of the school year, leaving the old and ushering in the new, in the golden autumn of September, ModelWhale ushered in a new round of version updates to continuously optimize your experience.
In this update, ModelWhale mainly performs the following functional iterations:
• New "teacher-student" dual-view switching (team version ✓)
• Add storage space application (team version ✓)
• New knowledge base search (Pro ✓ Team ✓)
• New Spatial Search (Pro ✓ Team ✓)
• Added model library search (Pro ✓ Team ✓)
• Added the use of supercomputing (HPC) to run Notebook (privatization)
• Apply the FAIR principle to the Whale community dataset
1 Added "teacher-student" dual-view switching (team version ✓)
Data science teaching needs to use specific case coding and explore real industry data to help students better understand theoretical methods and cultivate students' hands-on practical ability. ModelWhale integrates data science research tools, classroom teaching and homework management systems , and connects to the open source community to provide rich, high-quality practical cases, industry data, and competition training for students to form analysis ideas and broaden their horizons.
When using ModelWhale to build classes, teachers can not only call or upload rich teaching materials, teaching tools, and design various forms of homework (such as practical homework, group homework, automatic review, etc.), but also switch to "student homework" at any time . " Perspective: As a "student", check the courseware display, run the course code online, experience the submission process of various assignments, etc., to confirm that the teaching path design and teaching content display meet expectations. For more details, see: Teaching and Training Manual (Teacher) .
Tips: The teacher can also clear the homework submission records as a student with one click, and clear the impact of previous operations on the calculation of coursework and total grades.
2 Add storage space application (team version ✓)
ModelWhale provides each user with an independent disk space for file reading and writing and persistent storage , including input, project, temp, and work; Work is equivalent to each user's independent U disk (can be mounted to different code projects for use) : Organization administrators can perform unified allocation management , and organization members can apply for capacity expansion on demand .
When a member's work storage space is insufficient, an expansion application can be initiated: fill in the required expansion size and the reason for the application, and then submit it to the administrator for review; when the administrator reviews, he can choose to approve or reject, and can also quickly purchase the storage space for a fee.
Tips: (1) Before the application is approved (apply/rejected), members can modify the workspace capacity they want to apply for; (2) Members can also apply for additional computing resource time, and the application process is similar.
3 New knowledge base search (professional version ✓ team version ✓)
The ModelWhale knowledge base allows you to organize content across entities (including projects, data, files, videos, links, etc.) , making the accumulation of results more organized, and also supports creating folders and dragging and sorting. In addition to previewing data statistics, running code projects, and playing videos in the knowledge base, you can also search and filter content so that you can find the corresponding content more quickly.
4 Added spatial search (professional version ✓ team version ✓)
The ModelWhale space consists of "projects", "data" and "knowledge base" , and "knowledge base" can carry a variety of content entities, including projects, data, files, videos, links, etc. When the problems we want to solve become complex, it is inevitable that a lot of content will be generated in the process. Spatial content now also supports search and filtering, allowing you to find what you need more quickly.
5 Added model library search (professional version ✓ team version ✓)
The ModelWhale model library helps you manage the algorithms and models produced by you, and realizes the organization, sharing and reuse of them (we also organize some commonly used algorithm models for your invocation experience). The files in the model library support preview viewing , and now also support search and filtering , so that you can find what you need more quickly: (1) Search scope, including all "models created by myself" and "models shared by others";( 2) Filter automatically, including creator, framework (such as Keras, PyTorch, Tensorflow, MXNet), source model library, modification time. For more details, see: Model Development and Deployment .
6 Added the use of supercomputing (HPC) to run Notebook (privatization)
Many research tasks and complex computing tasks require the support of supercomputing (high-performance computing power). We support access to your existing high-performance computer cluster (HPC) to call computing power on the platform: the organization administrator completes Relevant configurations, and let users configure their own "Slurm Username" and "Slurm Token", they can call HPC on the platform to perform online Notebook computing, offline training task computing, and cloud hosting task computing.
Tips: This function is only open to privatization customers.
7 Applying the FAIR principle to the Whale Community Dataset
The Hejing community dataset has now applied the FAIR principle to support a more standardized and scientific dataset metadata display and increase data availability. All data science enthusiasts are welcome to come and explore .
The above is the whole content of the ModelWhale version update in this issue.
Enter ModelWhale.com , try the professional version (individual research) or the team version (organizational collaboration) for free, and get free CPU and GPU computing power! (It is recommended to use the computer for trial experience)
If you have any suggestions, questions about ModelWhale, or need to renew the trial, please contact MW , MoMo is happy to serve and communicate with you.