-
Notifications
You must be signed in to change notification settings - Fork 56
All issues
Issue creation is restricted in this repository
Issues
is:issue state:open
is:issue state:open
Search results
[Bug]: Wrong output with tp4 and granite 8b
bugSomething isn't workingSomething isn't workingStatus: Open.#1043 In torch-spyre/spyre-inference;[Bug]: TP4 inference hangs indefinitely on a lost collective completion (RuntimeStream::synchronize() still waiting, in_flight_ frozen)
bugSomething isn't workingSomething isn't workingStatus: Open.#1042 In torch-spyre/spyre-inference;- Status: Open.#1037 In torch-spyre/spyre-inference;
- Status: Open.#1033 In torch-spyre/spyre-inference;
[Feature]: Enable batched decode for Gemma-4
enhancementNew feature or requestNew feature or requestStatus: Open.#1025 In torch-spyre/spyre-inference;[Feature]: Add Model Registry configuration in spyre-inference
enhancementNew feature or requestNew feature or requestStatus: Open.[Bug]: vLLM segmentation fault due to incompatible protobuf version on s390x
bugSomething isn't workingSomething isn't workingStatus: Open.#1005 In torch-spyre/spyre-inference;- Status: Open.#985 In torch-spyre/spyre-inference;
Gemma 4 vision encoder: performance improvement
enhancementNew feature or requestNew feature or requestStatus: Open.#976 In torch-spyre/spyre-inference;[Feature]: confirm nothing depends on the 0-dim device index constraint
enhancementNew feature or requestNew feature or requestStatus: Open.#963 In torch-spyre/spyre-inference;[Feature]: revisit the host-side embedding pad now that padded compiled collectives are exact
enhancementNew feature or requestNew feature or requestStatus: Open.#962 In torch-spyre/spyre-inference;[Feature]: Continuously benchmark server preparation time on cold and populated cache
enhancementNew feature or requestNew feature or requestStatus: Open.#959 In torch-spyre/spyre-inference;