Open Issues Need Help
View All on GitHubA Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
AI Summary: The `logging::tests::test_json_log_capture` Rust test is intermittently failing after the merge of PR #2406. The test passes when run in isolation but fails when executed alongside other tests, indicating a potential test isolation issue or log pollution. The failure manifests as the test observing error logs from other parts of the system instead of its expected output.
A Datacenter Scale Distributed Inference Serving Framework
A Datacenter Scale Distributed Inference Serving Framework
AI Summary: The task is to identify and resolve the issue of multiple versions of the `candle-core` dependency being installed within the Dynamo project. This involves standardizing on a single version across the entire workspace to improve build times. The solution likely involves updating dependency specifications and resolving any resulting compatibility conflicts.
A Datacenter Scale Distributed Inference Serving Framework
AI Summary: Debug a bug in the NVIDIA Dynamo distributed inference serving framework where models appear in the model list before they are ready for inference, resulting in 500 errors on inference requests. The task involves analyzing the provided reproduction steps, understanding the Dynamo architecture (as described in the project README), and identifying the source of the timing discrepancy between model registration and readiness.
A Datacenter Scale Distributed Inference Serving Framework
AI Summary: Modify the Dynamo framework's mockers to avoid downloading model weights during configuration. This involves decoupling the mockers from the ModelDeploymentCard's weight download mechanism, allowing configuration to proceed using only the tokenizer information.
A Datacenter Scale Distributed Inference Serving Framework
AI Summary: Improve the error handling of the Dynamo LLM serving framework's `/chat/completions` endpoint. The current error message for requests with an empty `messages` field is cryptic; it needs to be replaced with a clear and informative message indicating that the `messages` field cannot be empty.
A Datacenter Scale Distributed Inference Serving Framework
AI Summary: Parallelize the tokenization process within the Dynamo framework's batch completion functionality using the Rayon library to improve performance. This involves identifying the appropriate sections of code for parallelization and integrating Rayon to handle the task efficiently.
A Datacenter Scale Distributed Inference Serving Framework