Skip to content

pullrequests Search Results · language:Dune language:Python language:Python language:JavaScript language:JavaScript

Filter by

149M results  (3 s)

149M results

Summary - Adds SubagentTokenUsageEvent emitted mid-stream whenever a subagent reports token usage - Includes input_tokens, output_tokens, cache_read_tokens, cache_creation_tokens, agent_id, session_id ...

Summary Auto-suggests a Posted Date for credit-card charges made on a non-business day (weekend or Ontario bank holiday), so transactions that won t settle until the next business day land in the correct ...

Summary - allow the hand-written TypeScript client inputs to accept the new observationScopes: shared retain option - keep the public helper type aligned with the generated client/OpenAPI union from ...

Description This pull request implements Step 10 of the frontend timeline: Code Quality and Final Documentation. This serves as the concluding PR for the Meta Glasses Reviews dashboard project, transitioning ...

snyk-top-banner Snyk has created this PR to fix 1 vulnerabilities in the maven dependencies of this project. Snyk changed the following file(s): - pom.xml Vulnerabilities that will be fixed with ...

This minor PR restores the launcher name (vLLM Name in the report) that was accidentally removed when the node name was being added as one of the columns. Having the launcher name makes it easier to correlate ...

…tarts PR template Select PR template in preview mode - [Hotfix](?expand=1 template=Hotfix_template.md) - [DA](?expand=1 template=DA_ticket_template.md) - [DE](?expand=1 template=DE_ticket_template.md) ...

Reverts ministryofjustice/prepare-a-case#1156

Summary Публичная лендинг-страница курса для лидогенерации (как у Qadam Education): Флоу 1. Рекламная ссылка /course/{courseId} → пользователь попадает на лендинг 2. Видит программу курса: модули, ...

Permet à N machines GPU réparties sur Internet d entraîner le même modèle en parallèle (data-parallel asynchrone façon FedAvg) avec un TRAINER_TOKEN partagé, sans s écraser : chaque trainer entraîne localement ...