topic Databricks GenAI & ML Announcements — December 2024 in Announcements

topic Databricks GenAI & ML Announcements — December 2024 in Announcements https://community.databricks.com/t5/announcements/databricks-genai-amp-ml-announcements-december-2024/m-p/105258#M229 <DIV class="eq er es et eu l"> <ARTICLE> <DIV class="l"> <DIV class="l"> <SECTION> <DIV> <DIV class="fj fk fl fm fn"> <DIV class="ab cb"> <DIV class="ci bh ev ew ex ey"> <H1 id="a37c" class="mj mk fq bf ml mm mn mo mp mq mr ms mt mu mv mw mx my mz na nb nc nd ne nf ng bk" data-selectable-paragraph="">Streamline AI Agent Evaluation using synthetic evaluation sets (Public Preview)</H1> <UL class=""> <LI id="3c66" class="la lb fq lc b ld nh lf lg lh ni lj lk ll nj ln lo lp nk lr ls lt nl lv lw lx nm nn no bk" data-selectable-paragraph="">Demo:<SPAN> </SPAN><A class="af mi" href="https://www.youtube.com/watch?v=8Mb91QtLzJ8" target="_blank" rel="noopener ugc nofollow">https://www.youtube.com/watch?v=8Mb91QtLzJ8</A></LI> <LI id="b479" class="la lb fq lc b ld np lf lg lh nq lj lk ll nr ln lo lp ns lr ls lt nt lv lw lx nm nn no bk" data-selectable-paragraph="">You can evaluate your AI agent by generating a representative evaluation set from your documents. The synthetic generation API is tightly integrated with Agent Evaluation, allowing you to quickly evaluate and improve the quality of your agent’s responses without going through the costly process of human labeling. See<SPAN> </SPAN><A class="af mi" href="https://docs.databricks.com/en/generative-ai/agent-evaluation/synthesize-evaluation-set.html" target="_blank" rel="noopener ugc nofollow">Synthesize evaluation sets</A>.</LI> <LI id="b204" class="la lb fq lc b ld np lf lg lh nq lj lk ll nr ln lo lp ns lr ls lt nt lv lw lx nm nn no bk" data-selectable-paragraph="">One effective method for synthesizing evaluation datasets involves using the<SPAN> </SPAN><CODE class="cx nu nv nw nx b">generate_evals_df</CODE><SPAN> </SPAN>method from the<SPAN> </SPAN><CODE class="cx nu nv nw nx b">databricks-agents</CODE><SPAN> </SPAN>Python package. This method requires a DataFrame with two columns:<SPAN> </SPAN><CODE class="cx nu nv nw nx b">content</CODE><SPAN> </SPAN>(the parsed document content as a string) and<SPAN> </SPAN><CODE class="cx nu nv nw nx b">doc_uri</CODE><SPAN> </SPAN>(the document's URI). Developers can control the generation process using key parameters:<SPAN> </SPAN><CODE class="cx nu nv nw nx b">num_evals</CODE><SPAN> </SPAN>(the total number of evaluations to generate),<SPAN> </SPAN><CODE class="cx nu nv nw nx b">agent_description</CODE><SPAN> </SPAN>(the task description of the agent), and<SPAN> </SPAN><CODE class="cx nu nv nw nx b">question_guidelines</CODE><SPAN> </SPAN>(guidelines for generating synthetic questions). The<SPAN> </SPAN><CODE class="cx nu nv nw nx b">num_evals</CODE><SPAN> </SPAN>parameter intelligently distributes evaluations across documents, balancing the number of questions per page and considering document size. The output includes detailed columns such as<SPAN> </SPAN><CODE class="cx nu nv nw nx b">request_id</CODE>,<SPAN> </SPAN><CODE class="cx nu nv nw nx b">request</CODE>,<SPAN> </SPAN><CODE class="cx nu nv nw nx b">expected_facts</CODE>, and<SPAN> </SPAN><CODE class="cx nu nv nw nx b">expected_retrieved_context</CODE>, ensuring traceability with the corresponding<SPAN> </SPAN><CODE class="cx nu nv nw nx b">doc_uri</CODE>. For developers unsure about how many evaluations are needed, the<SPAN> </SPAN><CODE class="cx nu nv nw nx b">estimate_synthetic_num_evals</CODE><SPAN> </SPAN>method helps estimate the ideal number of evaluations for desired coverage.</LI> <LI id="8a2c" class="la lb fq lc b ld np lf lg lh nq lj lk ll nr ln lo lp ns lr ls lt nt lv lw lx nm nn no bk" data-selectable-paragraph="">By using synthetic evaluation datasets, AI teams can automate and optimize the testing process, saving valuable time and achieving thorough validation of their agents’ capabilities.</LI> </UL> <H1 id="df6b" class="mj mk fq bf ml mm mn mo mp mq mr ms mt mu mv mw mx my mz na nb nc nd ne nf ng bk" data-selectable-paragraph="">Python code executor for AI agents (Public Preview)</H1> <UL class=""> <LI id="80ab" class="la lb fq lc b ld nh lf lg lh ni lj lk ll nj ln lo lp nk lr ls lt nl lv lw lx nm nn no bk" data-selectable-paragraph="">You can now quickly give your AI agents the ability to run Python code. Databricks now offers a pre-built Unity Catalog function that can be used by an AI agent as a tool to expand their capabilities beyond language generation. See<SPAN> </SPAN><A class="af mi" href="https://docs.databricks.com/en/generative-ai/agent-framework/code-interpreter-tools.html" target="_blank" rel="noopener ugc nofollow">Code interpreter AI agent tools</A>.</LI> </UL> <H1 id="9861" class="mj mk fq bf ml mm mn mo mp mq mr ms mt mu mv mw mx my mz na nb nc nd ne nf ng bk" data-selectable-paragraph="">Add budget policies to model serving endpoints (Public Preview)</H1> <UL class=""> <LI id="c48c" class="la lb fq lc b ld nh lf lg lh ni lj lk ll nj ln lo lp nk lr ls lt nl lv lw lx nm nn no bk" data-selectable-paragraph="">Demo:<SPAN> </SPAN><A class="af mi" href="https://www.youtube.com/watch?v=6y9rpReGquM" target="_blank" rel="noopener ugc nofollow">https://www.youtube.com/watch?v=6y9rpReGquM</A></LI> <LI id="e94a" class="la lb fq lc b ld np lf lg lh nq lj lk ll nr ln lo lp ns lr ls lt nt lv lw lx nm nn no bk" data-selectable-paragraph="">Effectively managing cloud computing costs is essential, especially when working with serverless compute resources. Databricks addresses this need with<SPAN> </SPAN><A class="af mi" href="https://docs.databricks.com/en/admin/usage/budget-policies.html" target="_blank" rel="noopener ugc nofollow"><STRONG class="lc fr">budget policies</STRONG></A>, which help organizations track and control serverless usage. These policies work by applying tags to any serverless compute activity associated with the policy. These tags are logged in billing records, making it easier to attribute serverless spending to specific budgets and perform granular billing analysis. A notable update is that budget policies are now supported on<SPAN> </SPAN><STRONG class="lc fr">model serving endpoints</STRONG>, enabling organizations to monitor and control serverless costs tied to machine learning models. See<SPAN> </SPAN><A class="af mi" href="https://docs.databricks.com/en/machine-learning/model-serving/manage-serving-endpoints.html" target="_blank" rel="noopener ugc nofollow">Manage model serving endpoints</A>.</LI> <LI id="22be" class="la lb fq lc b ld np lf lg lh nq lj lk ll nr ln lo lp ns lr ls lt nt lv lw lx nm nn no bk" data-selectable-paragraph="">To get started, workspace admins can create budgets in the<SPAN> </SPAN><STRONG class="lc fr">Account Console</STRONG><SPAN> </SPAN>and apply budget policies directly through the Databricks UI. Admins can also manage and view policies they have created or have permissions for. However, to manage all policies across an account, admins must also have the<SPAN> </SPAN><STRONG class="lc fr">Billing admin</STRONG><SPAN> </SPAN>role at the account level. Non-admin users can manage budget policies if they are assigned the<SPAN> </SPAN><STRONG class="lc fr">Budget Policy Manager</STRONG><SPAN> </SPAN>permissions.</LI> <LI id="ac35" class="la lb fq lc b ld np lf lg lh nq lj lk ll nr ln lo lp ns lr ls lt nt lv lw lx nm nn no bk" data-selectable-paragraph="">Applying a budget policy is straightforward. When creating a model serving endpoint, users can select a budget policy from the<SPAN> </SPAN><STRONG class="lc fr">Budget Policy</STRONG><SPAN> </SPAN>menu in the Serving UI. If a budget policy is already assigned to a user, it will automatically apply to any new endpoints they create. Existing endpoints, however, need to be manually updated to include a budget policy. Additionally, users with<SPAN> </SPAN><STRONG class="lc fr">MANAGE</STRONG><SPAN> </SPAN>permissions can modify budget policies for existing endpoints via the<SPAN> </SPAN><STRONG class="lc fr">Endpoint Details</STRONG><SPAN> </SPAN>page.</LI> <LI id="e127" class="la lb fq lc b ld np lf lg lh nq lj lk ll nr ln lo lp ns lr ls lt nt lv lw lx nm nn no bk" data-selectable-paragraph="">It’s important to note that this feature is currently in<SPAN> </SPAN><STRONG class="lc fr">Public Preview</STRONG><SPAN> </SPAN>and does not support endpoints serving<SPAN> </SPAN><STRONG class="lc fr">External Models</STRONG><SPAN> </SPAN>or<SPAN> </SPAN><STRONG class="lc fr">Foundation Model APIs</STRONG><SPAN> </SPAN>with pay-per-token workloads. Moreover, existing endpoints won’t automatically inherit newly assigned budget policies — they must be manually updated.</LI> <LI id="0836" class="la lb fq lc b ld np lf lg lh nq lj lk ll nr ln lo lp ns lr ls lt nt lv lw lx nm nn no bk" data-selectable-paragraph="">By implementing budget policies, organizations can achieve greater visibility and control over their serverless spending, ensuring more effective resource management and cost optimization.</LI> </UL> <FIGURE class="ly lz ma mb mc md ny nz paragraph-image"> <DIV class="ob oc ed od bh oe" tabindex="0" role="button"> <DIV class="ny nz oa"><PICTURE><SOURCE srcset="https://miro.medium.com/v2/resize:fit:640/format:webp/0*KNBOSyeOremrD2bi 640w, https://miro.medium.com/v2/resize:fit:720/format:webp/0*KNBOSyeOremrD2bi 720w, https://miro.medium.com/v2/resize:fit:750/format:webp/0*KNBOSyeOremrD2bi 750w, https://miro.medium.com/v2/resize:fit:786/format:webp/0*KNBOSyeOremrD2bi 786w, https://miro.medium.com/v2/resize:fit:828/format:webp/0*KNBOSyeOremrD2bi 828w, https://miro.medium.com/v2/resize:fit:1100/format:webp/0*KNBOSyeOremrD2bi 1100w, https://miro.medium.com/v2/resize:fit:1400/format:webp/0*KNBOSyeOremrD2bi 1400w" type="image/webp" sizes="(min-resolution: 4dppx) and (max-width: 700px) 50vw, (-webkit-min-device-pixel-ratio: 4) and (max-width: 700px) 50vw, (min-resolution: 3dppx) and (max-width: 700px) 67vw, (-webkit-min-device-pixel-ratio: 3) and (max-width: 700px) 65vw, (min-resolution: 2.5dppx) and (max-width: 700px) 80vw, (-webkit-min-device-pixel-ratio: 2.5) and (max-width: 700px) 80vw, (min-resolution: 2dppx) and (max-width: 700px) 100vw, (-webkit-min-device-pixel-ratio: 2) and (max-width: 700px) 100vw, 700px"></SOURCE><SOURCE srcset="https://miro.medium.com/v2/resize:fit:640/0*KNBOSyeOremrD2bi 640w, https://miro.medium.com/v2/resize:fit:720/0*KNBOSyeOremrD2bi 720w, https://miro.medium.com/v2/resize:fit:750/0*KNBOSyeOremrD2bi 750w, https://miro.medium.com/v2/resize:fit:786/0*KNBOSyeOremrD2bi 786w, https://miro.medium.com/v2/resize:fit:828/0*KNBOSyeOremrD2bi 828w, https://miro.medium.com/v2/resize:fit:1100/0*KNBOSyeOremrD2bi 1100w, https://miro.medium.com/v2/resize:fit:1400/0*KNBOSyeOremrD2bi 1400w" sizes="(min-resolution: 4dppx) and (max-width: 700px) 50vw, (-webkit-min-device-pixel-ratio: 4) and (max-width: 700px) 50vw, (min-resolution: 3dppx) and (max-width: 700px) 67vw, (-webkit-min-device-pixel-ratio: 3) and (max-width: 700px) 65vw, (min-resolution: 2.5dppx) and (max-width: 700px) 80vw, (-webkit-min-device-pixel-ratio: 2.5) and (max-width: 700px) 80vw, (min-resolution: 2dppx) and (max-width: 700px) 100vw, (-webkit-min-device-pixel-ratio: 2) and (max-width: 700px) 100vw, 700px" data-testid="og"></SOURCE></PICTURE><span class="lia-inline-image-display-wrapper lia-image-align-inline" image-alt="lara_rachidi_0-1736538849055.png" style="width: 400px;"><img src="https://community.databricks.com/t5/image/serverpage/image-id/14052iA36BB40504172C86/image-size/medium?v=v2&px=400" role="button" title="lara_rachidi_0-1736538849055.png" alt="lara_rachidi_0-1736538849055.png" /></span> <P> </P> </DIV> </DIV> </FIGURE> <H1 id="aa4f" class="mj mk fq bf ml mm mn mo mp mq mr ms mt mu mv mw mx my mz na nb nc nd ne nf ng bk" data-selectable-paragraph="">Mosaic AI Model Training serverless forecasting (Public Preview)</H1> <UL class=""> <LI id="8d59" class="la lb fq lc b ld nh lf lg lh ni lj lk ll nj ln lo lp nk lr ls lt nl lv lw lx nm nn no bk" data-selectable-paragraph="">Demo:<SPAN> </SPAN><A class="af mi" href="https://www.youtube.com/watch?v=WXF88P7tHCA" target="_blank" rel="noopener ugc nofollow">https://www.youtube.com/watch?v=WXF88P7tHCA</A></LI> <LI id="5ad9" class="la lb fq lc b ld np lf lg lh nq lj lk ll nr ln lo lp ns lr ls lt nt lv lw lx nm nn no bk" data-selectable-paragraph=""><A class="af mi" href="https://docs.databricks.com/en/machine-learning/train-model/serverless-forecasting.html" target="_blank" rel="noopener ugc nofollow">Mosaic AI Model Training — forecasting</A><SPAN> </SPAN>improves upon the existing AutoML forecasting experience with serverless compute, Unity Catalog support, access to deep learning algorithms, and an upgraded interface.</LI> <LI id="b24c" class="la lb fq lc b ld np lf lg lh nq lj lk ll nr ln lo lp ns lr ls lt nt lv lw lx nm nn no bk" data-selectable-paragraph="">The goal is to simplify time-series forecasting by automatically selecting the optimal algorithms and hyperparameters while running on fully managed, scalable compute resources.</LI> <LI id="9392" class="la lb fq lc b ld np lf lg lh nq lj lk ll nr ln lo lp ns lr ls lt nt lv lw lx nm nn no bk" data-selectable-paragraph=""><STRONG class="lc fr">Getting Started with Serverless Forecasting</STRONG><SPAN> </SPAN>is straightforward. All you need is a training dataset with a time-series column stored as a<SPAN> </SPAN><STRONG class="lc fr">Unity Catalog table</STRONG>. If your workspace uses a<SPAN> </SPAN><STRONG class="lc fr">Secure Egress Gateway (SEG)</STRONG>, make sure to add<SPAN> </SPAN><CODE class="cx nu nv nw nx b">pypi.org</CODE><SPAN> </SPAN>to the allowed domains list to avoid connectivity issues. This setup allows seamless forecasting without worrying about infrastructure management.</LI> <LI id="5346" class="la lb fq lc b ld np lf lg lh nq lj lk ll nr ln lo lp ns lr ls lt nt lv lw lx nm nn no bk" data-selectable-paragraph="">To begin, navigate to the<SPAN> </SPAN><STRONG class="lc fr">Experiments</STRONG><SPAN> </SPAN>tab in Databricks and use the sample dataset<SPAN> </SPAN><CODE class="cx nu nv nw nx b">dbdemos.dbdemos_iot_turbine.turbine_training_dataset_ml</CODE>. Set the forecast frequency (e.g., hourly) and follow these steps:</LI> <LI id="bb3d" class="la lb fq lc b ld np lf lg lh nq lj lk ll nr ln lo lp ns lr ls lt nt lv lw lx nm nn no bk" data-selectable-paragraph=""><STRONG class="lc fr">Select Training Data:</STRONG><SPAN> </SPAN>Choose your dataset from accessible Unity Catalog tables.</LI> <LI id="835e" class="la lb fq lc b ld np lf lg lh nq lj lk ll nr ln lo lp ns lr ls lt nt lv lw lx nm nn no bk" data-selectable-paragraph=""><STRONG class="lc fr">Configure Columns:</STRONG></LI> <LI id="c972" class="la lb fq lc b ld np lf lg lh nq lj lk ll nr ln lo lp ns lr ls lt nt lv lw lx nm nn no bk" data-selectable-paragraph=""><STRONG class="lc fr">Time Column:</STRONG><SPAN> </SPAN>Identify the column with timestamps or dates.</LI> <LI id="e6d5" class="la lb fq lc b ld np lf lg lh nq lj lk ll nr ln lo lp ns lr ls lt nt lv lw lx nm nn no bk" data-selectable-paragraph=""><STRONG class="lc fr">Forecast Frequency:</STRONG><SPAN> </SPAN>Define the data’s time unit (minutes, hours, days, months).</LI> <LI id="549e" class="la lb fq lc b ld np lf lg lh nq lj lk ll nr ln lo lp ns lr ls lt nt lv lw lx nm nn no bk" data-selectable-paragraph=""><STRONG class="lc fr">Forecast Horizon:</STRONG><SPAN> </SPAN>Specify how far into the future to predict.</LI> <LI id="7b5d" class="la lb fq lc b ld np lf lg lh nq lj lk ll nr ln lo lp ns lr ls lt nt lv lw lx nm nn no bk" data-selectable-paragraph=""><STRONG class="lc fr">Prediction Target Column:</STRONG><SPAN> </SPAN>Choose the feature to forecast.</LI> <LI id="b781" class="la lb fq lc b ld np lf lg lh nq lj lk ll nr ln lo lp ns lr ls lt nt lv lw lx nm nn no bk" data-selectable-paragraph=""><STRONG class="lc fr">Optional Settings:</STRONG></LI> <LI id="f02d" class="la lb fq lc b ld np lf lg lh nq lj lk ll nr ln lo lp ns lr ls lt nt lv lw lx nm nn no bk" data-selectable-paragraph=""><STRONG class="lc fr">Prediction Data Path:</STRONG><SPAN> </SPAN>Store output forecasts in a Unity Catalog table.</LI> <LI id="2806" class="la lb fq lc b ld np lf lg lh nq lj lk ll nr ln lo lp ns lr ls lt nt lv lw lx nm nn no bk" data-selectable-paragraph=""><STRONG class="lc fr">Model Registration Location:</STRONG><SPAN> </SPAN>Define where to save the trained model.</LI> <LI id="4cac" class="la lb fq lc b ld np lf lg lh nq lj lk ll nr ln lo lp ns lr ls lt nt lv lw lx nm nn no bk" data-selectable-paragraph=""><STRONG class="lc fr">Advanced Options:</STRONG><SPAN> </SPAN>Customize experiment names, identifier columns for multi-series forecasting, evaluation metrics, training frameworks, data splits, weighting, holiday regions, and timeout settings.</LI> </UL> <P class="pw-post-body-paragraph la lb fq lc b ld le lf lg lh li lj lk ll lm ln lo lp lq lr ls lt lu lv lw lx fj bk" data-selectable-paragraph="">Once configured, click<SPAN> </SPAN><STRONG class="lc fr">Start Training</STRONG><SPAN> </SPAN>to run the AutoML experiment. You can monitor progress, stop the experiment if needed, and explore results in real time. After training, the forecast results are saved in a Delta table, and the best-performing model is registered in the Unity Catalog.</P> <P class="pw-post-body-paragraph la lb fq lc b ld le lf lg lh li lj lk ll lm ln lo lp lq lr ls lt lu lv lw lx fj bk" data-selectable-paragraph="">From there, you can:</P> <UL class=""> <LI id="7749" class="la lb fq lc b ld le lf lg lh li lj lk ll lm ln lo lp lq lr ls lt lu lv lw lx nm nn no bk" data-selectable-paragraph=""><STRONG class="lc fr">View Predictions:</STRONG><SPAN> </SPAN>Examine the forecasting results table.</LI> <LI id="17e4" class="la lb fq lc b ld np lf lg lh nq lj lk ll nr ln lo lp ns lr ls lt nt lv lw lx nm nn no bk" data-selectable-paragraph=""><STRONG class="lc fr">Run Batch Inference:</STRONG><SPAN> </SPAN>Use the auto-generated notebook for batch predictions.</LI> <LI id="0d1c" class="la lb fq lc b ld np lf lg lh nq lj lk ll nr ln lo lp ns lr ls lt nt lv lw lx nm nn no bk" data-selectable-paragraph=""><STRONG class="lc fr">Deploy the Model:</STRONG><SPAN> </SPAN>Easily create a serving endpoint for real-time forecasting.</LI> </UL> <P class="pw-post-body-paragraph la lb fq lc b ld le lf lg lh li lj lk ll lm ln lo lp lq lr ls lt lu lv lw lx fj bk" data-selectable-paragraph="">For a detailed comparison between<SPAN> </SPAN><STRONG class="lc fr">serverless forecasting</STRONG><SPAN> </SPAN>and traditional<SPAN> </SPAN><STRONG class="lc fr">classic compute forecasting</STRONG>, check out the official documentation<SPAN> </SPAN><A class="af mi" href="https://docs.databricks.com/en/machine-learning/train-model/serverless-forecasting.html#serverless-forecasting-vs-classic-compute-forecasting" target="_blank" rel="noopener ugc nofollow">here</A>.</P> <P class="pw-post-body-paragraph la lb fq lc b ld le lf lg lh li lj lk ll lm ln lo lp lq lr ls lt lu lv lw lx fj bk" data-selectable-paragraph="">This serverless forecasting feature empowers data teams to quickly build and deploy accurate time-series models without the burden of managing infrastructure, making forecasting more accessible and efficient than ever before.</P> <H1 id="8c96" class="mj mk fq bf ml mm mn mo mp mq mr ms mt mu mv mw mx my mz na nb nc nd ne nf ng bk" data-selectable-paragraph="">Platform Updates</H1> <UL class=""> <LI id="d97b" class="la lb fq lc b ld nh lf lg lh ni lj lk ll nj ln lo lp nk lr ls lt nl lv lw lx nm nn no bk" data-selectable-paragraph=""><A class="af mi" href="https://docs.databricks.com/en/release-notes/product/2024/december.html#databricks-agents-sdk-0130-release" target="_blank" rel="noopener ugc nofollow">databricks-agents SDK 0.13.0 release</A>: Version 0.13.0 of the<SPAN> </SPAN><CODE class="cx nu nv nw nx b">databricks-agents</CODE><SPAN> </SPAN>SDK has been<SPAN> </SPAN><A class="af mi" href="https://pypi.org/project/databricks-agents/" target="_blank" rel="noopener ugc nofollow">released to PyPI</A>.</LI> <LI id="4a6a" class="la lb fq lc b ld np lf lg lh nq lj lk ll nr ln lo lp ns lr ls lt nt lv lw lx nm nn no bk" data-selectable-paragraph=""><A class="af mi" href="https://docs.databricks.com/en/release-notes/product/2024/december.html#meta-llama-33-is-now-available-for-provisioned-throughput-workloads" target="_blank" rel="noopener ugc nofollow">Meta Llama 3.3 is now available for provisioned throughput workloads</A></LI> <LI id="39df" class="la lb fq lc b ld np lf lg lh nq lj lk ll nr ln lo lp ns lr ls lt nt lv lw lx nm nn no bk" data-selectable-paragraph=""><A class="af mi" href="https://docs.databricks.com/en/release-notes/product/2024/december.html#meta-llama-33-70b-instruct-is-now-available-on-model-serving" target="_blank" rel="noopener ugc nofollow">Meta Llama 3.3 70B Instruct is now available on Model Serving</A></LI> <LI id="fa3a" class="la lb fq lc b ld np lf lg lh nq lj lk ll nr ln lo lp ns lr ls lt nt lv lw lx nm nn no bk" data-selectable-paragraph=""><A class="af mi" href="https://docs.databricks.com/en/release-notes/product/2024/december.html#bamboolib-is-now-deprecated" target="_blank" rel="noopener ugc nofollow">bamboolib is now deprecated</A></LI> </UL> <H1 id="b384" class="mj mk fq bf ml mm mn mo mp mq mr ms mt mu mv mw mx my mz na nb nc nd ne nf ng bk" data-selectable-paragraph="">Blog Posts</H1> <H2 id="3255" class="og mk fq bf ml oh oi oj mp ok ol om mt ll on oo op lp oq or os lt ot ou ov ow bk" data-selectable-paragraph="">Benchmarking Domain Intelligence</H2> <UL class=""> <LI id="4456" class="la lb fq lc b ld nh lf lg lh ni lj lk ll nj ln lo lp nk lr ls lt nl lv lw lx nm nn no bk" data-selectable-paragraph="">Blog post available<SPAN> </SPAN><A class="af mi" href="https://www.databricks.com/blog/benchmarking-domain-intelligence" target="_blank" rel="noopener ugc nofollow">here</A>.</LI> </UL> <H2 id="fa2a" class="og mk fq bf ml oh oi oj mp ok ol om mt ll on oo op lp oq or os lt ot ou ov ow bk" data-selectable-paragraph="">Batch Inference on Fine Tuned Llama Models with Mosaic AI Model Serving</H2> <UL class=""> <LI id="c13c" class="la lb fq lc b ld nh lf lg lh ni lj lk ll nj ln lo lp nk lr ls lt nl lv lw lx nm nn no bk" data-selectable-paragraph="">Blog post available<SPAN> </SPAN><A class="af mi" href="https://www.databricks.com/blog/batch-inference-fine-tuned-llama-models-mosaic-ai-model-serving" target="_blank" rel="noopener ugc nofollow">here</A>.</LI> </UL> <H2 id="5119" class="og mk fq bf ml oh oi oj mp ok ol om mt ll on oo op lp oq or os lt ot ou ov ow bk" data-selectable-paragraph="">Build an Autonomous AI Assistant with Mosaic AI Agent Framework</H2> <UL class=""> <LI id="cfd4" class="la lb fq lc b ld nh lf lg lh ni lj lk ll nj ln lo lp nk lr ls lt nl lv lw lx nm nn no bk" data-selectable-paragraph="">Blog post available<SPAN> </SPAN><A class="af mi" href="https://www.databricks.com/blog/build-autonomous-ai-assistant-mosaic-ai-agent-framework" target="_blank" rel="noopener ugc nofollow">here</A>.</LI> </UL> <H2 id="9320" class="og mk fq bf ml oh oi oj mp ok ol om mt ll on oo op lp oq or os lt ot ou ov ow bk" data-selectable-paragraph="">Aimpoint Digital: AI Agent Systems for Building Travel Itineraries</H2> <UL class=""> <LI id="eb16" class="la lb fq lc b ld nh lf lg lh ni lj lk ll nj ln lo lp nk lr ls lt nl lv lw lx nm nn no bk" data-selectable-paragraph="">Blog post available<SPAN> </SPAN><A class="af mi" href="https://www.databricks.com/blog/aimpoint-digital-ai-agent-systems" target="_blank" rel="noopener ugc nofollow">here</A>.</LI> </UL> </DIV> </DIV> </DIV> </DIV> </SECTION> </DIV> </DIV> </ARTICLE> </DIV> <DIV class="l"> </DIV> <FOOTER class="ox oy oz pa pb ab q pc pd c"> <DIV class="l ae"> <DIV class="ab cb"> <DIV class="ci bh ev ew ex ey"> <DIV class="ab cp pe"> <DIV class="ab q iu"> <DIV class="pf l"> <DIV class="ab q iu iv"> <DIV class="pw-multi-vote-icon ed xj ix iy iz"> <DIV class=""> <DIV> <DIV class="bm" aria-hidden="false" aria-describedby="64" aria-labelledby="64"> <DIV class="ja xk jc jd je jf jg am jh ji jj iz"> </DIV> </DIV> </DIV> </DIV> </DIV> </DIV> </DIV> <DIV class="bq ab"> <DIV> <DIV class="bm" aria-hidden="false" aria-describedby="8" aria-labelledby="8"> </DIV> </DIV> </DIV> </DIV> <DIV class="ab q"> <DIV class="pl l hi"> <DIV> <DIV class="bm" aria-hidden="false" aria-describedby="9" aria-labelledby="9"> <DIV class="bm" aria-hidden="false"> </DIV> </DIV> </DIV> </DIV> <DIV class="pl l hi"> <DIV class="bm" aria-hidden="false" aria-describedby="postFooterSocialMenu" aria-labelledby="postFooterSocialMenu"> <DIV> <DIV class="bm" aria-hidden="false" aria-describedby="10" aria-labelledby="10"> </DIV> </DIV> </DIV> </DIV> <DIV class="bm" aria-hidden="false"> <DIV class="bm" aria-hidden="false" aria-describedby="creatorActionOverflowMenu" aria-labelledby="creatorActionOverflowMenu"> <DIV class="sj l hi"> <DIV> <DIV class="bm" aria-hidden="false" aria-describedby="189" aria-labelledby="189"> </DIV> </DIV> </DIV> </DIV> </DIV> </DIV> </DIV> </DIV> </DIV> </DIV> </FOOTER> <DIV class="pm l"> <DIV class="ab cb"> <DIV class="ci bh ev ew ex ey"> <DIV class="ab pn po pp hk hj"> <DIV class="pq pr ps pt pu pv pw px py pz ab cp"> <DIV class="h k"> </DIV> </DIV> </DIV> </DIV> </DIV> </DIV> Fri, 10 Jan 2025 19:54:20 GMT lara_rachidi 2025-01-10T19:54:20Z Databricks GenAI & ML Announcements — December 2024 https://community.databricks.com/t5/announcements/databricks-genai-amp-ml-announcements-december-2024/m-p/105258#M229 <P><SPAN>→ Subscribe to our </SPAN><A class="af mi" href="https://www.youtube.com/channel/UCIRWm64xQdLY_0DO8omKHpw" target="_blank" rel="noopener ugc nofollow"><STRONG class="lc fr">YouTube channel</STRONG></A><SPAN> for regular updates</SPAN></P> Fri, 10 Jan 2025 19:54:20 GMT https://community.databricks.com/t5/announcements/databricks-genai-amp-ml-announcements-december-2024/m-p/105258#M229 lara_rachidi 2025-01-10T19:54:20Z